Artificial insemination and parthenogenesis in the whitespotted bamboo shark Chiloscyllium plagiosum

Non-lethal methods for semen collection from elasmobranchs to better understand species reproduction has accompanied the development of artificial insemination. Ejaculates (n = 82) collected from whitespotted bamboo sharks Chiloscyllium plagiosum (n = 19) were assessed and cold-stored raw or extended at 4 °C. Females (n = 20) were inseminated with fresh or 24–48 h cold-stored raw or extended semen and paternity of offspring determined with microsatellite markers. Insemination of females with fresh semen (n = 10) resulted in 80 hatchlings and 27.6% fertility. Insemination of females with semen cold-stored 24 h (n = 4) and 48 h (n = 1) semen resulted in 17 hatchlings and fertilization rates of 28.1% and 7.1% respectively. Two females inseminated with fresh or cold-stored semen laid eggs that hatched from fertilization and parthenogenesis within the same clutch. Parthenogenesis rate for inseminated females was 0.71%. Results demonstrate artificial insemination with cold-stored semen can provide a strategy for transport of male genetics nationally and internationally, precluding the need to transport sharks. Production of parthenotes in the same clutch as sexually fertilized eggs highlights the prevalence of parthenogenesis in whitespotted bamboo sharks and poses important considerations for population management.

www.nature.com/scientificreports/ short-term preservation of fish sperm and eggs for hours to weeks with the conditions of storage such as antibiotics and diluents optimized for each species [23][24][25][26][27][28] . Gamete cold storage does not require specialized equipment and could be useful in aquariums to allow the movement of genetic material between institutions to maintain gene diversity in a population without the need to transport an animal. This is especially useful as cryopreservation research using elasmobranch semen is limited to a few species of stingray 9,15 and investigation into cryopreservation of semen from whitespotted bamboo sharks has not yielded viable sperm (Wyffels, unpublished). The ability of elasmobranchs to reproduce by parthenogenesis is widely known for sharks (swell shark Cephaloscyllium ventriosum 29 ; bonnethead shark Sphyrna tiburo 30 ; blacktip shark Carcharinus limbatus 31 ; whitespotted bamboo C. plagiosum 7 and zebra sharks S. tigrinum 32,33 ) and rays (spotted eagle ray Aetobatus narinari 34 ), and documented for wild smalltooth sawfish Prisitis pectinata 35 . The strategy behind this form of reproduction is unknown, but it may be a reproductive adaptation for females after prolonged periods without male interaction. As such, it is assumed to occur instead of fertilization from a male. From a population management perspective, parthenotes do not contribute to the genetic diversity of a population and are rather female 'clones' of the dam. The overall incidence of parthenogenesis in whitespotted bamboo sharks is currently unknown, but realizing the incidence of parthenogenesis would aid with population modeling. Microsatellite markers have been developed previously to examine parentage in the whitespotted bamboo shark 7,43 .
The objectives of this study were to: (1) develop a safe and reliable method for non-lethal collection and characterization of semen from whitespotted bamboo sharks; (2) examine the effects of cold storage on sperm longevity; (3) produce offspring following AI with fresh and cold-stored semen; (4) develop additional microsatellite markers to confirm parentage of AI offspring; and (5) determine the incidence of parthenogenesis in whitespotted bamboo sharks.

Results
Ejaculate characteristics. Ejaculate collection was explored using two techniques (catheterization and manual expression) and collection by manual expression became the default collection method because of its relative simplicity and efficiency. A total of 82 ejaculates from 19 sharks were collected all months of the year attempted and semen characteristics are presented in Table 1. Ejaculates were light brown or buff colored, opaque and viscous containing individual sperm and spermatozeugmata composed of radially arrayed sperm with heads that were closely aligned and embedded in a matrix core (Fig. 1a) 21 . Spermatozeugmata were observed in 48% of ejaculates irrespective of time of year. Spermatozoa had elongated helical heads ( Fig. 1c) with an apical acrosome that was stained by Arachis hypogaea lectin or peanut agglutinin (PNA) (Fig. 1b-e). Sperm morphology was 100% normal for all ejaculates. The length of the head, midpiece and flagellum as well as the number of gyres for spermatozoa are shown in Table 2.
Spermatozoa in raw ejaculates had little or no motility but acquired motility after dilution in artificial seawater (ASW). After dilution in ASW, the cytoplasmic sleeve, found at the junction of the midpiece and tail (Fig. 1c), was observed to slide toward the distal end of the flagellum until it was shed from the spermatozoa. Shed cytoplasmic sleeves accumulated at the peripheral margin of the spermatozeugmata. Individual, free spermatozoa moving progressively had serpentine tails and demonstrated rotation about the long axis.
Short-term semen storage. Sample preparation method (p < 0.05) and duration of cold storage (p < 0.05) affected plasma membrane integrity (PMI) and total motility with diluted samples and shorter storage times resulting in higher PMI and total motility (Fig. 2). There was a strong positive correlation between sperm PMI and total motility (r = 0.93, n = 82, p < 0.001). There were no observed differences in PMI or total motility for semen extended in either ASW or elasmobranch adjusted Hank's balanced salt solution (E-HBSS), both of which supported sperm motility and PMI for up to 18 days (Fig. 2). Semen stored raw retained PMI, however was completely immotile by day 18. Cold storage of raw ejaculate for 24 h did not affect PMI (p = 0.268) but total motility decreased (p = 0.009). For semen cold-stored for 2-7 days, ejaculates extended in ASW or E-HBSS had higher PMI and motility than samples stored raw (p < 0.05, Fig. 2). Samples stored raw declined in quality the fastest of the three preparation methods and had the lowest PMI and total motility each day samples were evaluated (Fig. 2). 10 min and no adverse effects were observed. If egg cases were present in the oviduct the procedure was more challenging but the egg cases are leathery and durable and none were physically affected by the AI procedure, as evidenced by all egg cases laid being without damage. No eggs laid during the six weeks prior to AI developed. Fifteen females laid 114 fertile eggs resulting in 97 hatchlings after insemination and there was considerable variation in the number of eggs laid , duration or number of days of fertility (1-107), fertility (4.2-66.7%) and hatchability (2.6-60.0%) between females receiving 100-750 × 10 6 spermatozoa in raw and extended inseminates. Five females laid eggs (n = 66) 1-153 days after AI but none were fertile (Table 3). For most females (n = 9), all fertile eggs completed development successfully, but three females laid one fertile egg, two females laid two fertile eggs, and one female laid seven fertile eggs that all died before hatching (Supplemental Table 1). Average (± standard deviation) incubation was 117 ± 10 days and ranged from 74-142 days (n = 73). Successful insemination of both oviducts was confirmed by observation of development of eggs laid as pairs and was observed for five females that incidentally also had the highest fertility (37.5-66.7%) and hatchability (34.4-60.0%) rates (Table 3). Bilateral insemination success was not associated with inseminate age (fresh or 24 h), preparation    www.nature.com/scientificreports/ (raw or extended), or sperm dose (100-750 × 10 6 total sperm) and therefore possibly related to the technique. For the remainder of the females in successful trials (n = 10), only one of a pair of egg cases laid together was fertile (Fig. 3).
The combined average number of egg cases laid per females in successful trials was 24, double the combined average number of egg cases laid by females in unsuccessful trials, 12. The average fertility duration for females that produced nine or more hatchlings (n = 6) was 88 ± 9 days compared to 45 ± 5 days for females that produced 2-5 hatchlings (n = 4) ( Table 3). All eggs laid for 43 days after AI were fertile for female (ID 101) and one egg hatched from the first pair of egg cases laid after AI by four females on day six, eight or 13 post-insemination ( Fig. 3). Remaining females (n = 11) laid 1-10 infertile eggs before their first fertilized egg. Although females continued to lay fertile eggs for up to 121 days after insemination, fertility peaked by four weeks and decreased incrementally after 10 weeks (Fig. 4). All females that laid one or more fertile eggs also laid multiple infertile eggs afterwards, at the end of their laying cycle. Two females laid a single fertile egg 10 or 13 days post-insemination but not afterwards. Time from AI to laying date for the first fertilized egg ranged from 6-60 days and females laid fertile eggs 6-121 days after AI for 1-107 days (Table 3). There was no difference in fertility for ejaculates that were split and inseminated as raw semen or semen extended 1:1 with ASW (paired t-test: t = -0.08, df = 3, p = 0.9424). Collectively, nine females laid 61 fertilized eggs after AI using raw semen compared to 50 fertilized eggs laid by six females inseminated with extended semen (Supplemental Table 1).
Inseminates varied in sperm number from 3-750 × 10 6 spermatozoa and fertile eggs resulted from all doses except the lowest dose, 3 × 10 6 spermatozoa. However, only one of the two females that received the lowest dose laid eggs regularly post-insemination and can be considered in analyses. Insemination using 100-750 × 10 6 spermatozoa was effective and there was high variability in the number of hatchlings produced among the females even when using semen from the same male. For example, among 10 females inseminated with fresh semen from male 381, seven females produced 1-18 hatchlings each and three females produced no hatchlings (Supplemental Table 1).
Cold-stored semen successfully fertilized eggs for up to 48 h after collection and inseminates from 24 h coldstored semen had similar fertility and hatchability as inseminates from fresh semen. The same ejaculate split and used fresh and after 48 h of cold storage both fertilized eggs with three hatchlings from the fresh inseminate and one from the cold-stored inseminate. Two males contributed semen that was cold-stored for 24 h and shipped overnight between aquaria before being used for insemination, and both ejaculates fertilized eggs. Semen collected from male 395 at The Florida Aquarium and shipped overnight for AI trials at Ripley's Aquarium of the Smokies was used to inseminate three females that resulted in a total of 14 hatchlings. Semen from Adventure Aquarium male 2969 was shipped to Aquarium of the Pacific and used to inseminate a female that laid two fertile eggs that successfully hatched. Combined, 17 hatchlings resulted from AI of five females with cold-stored semen from three males from three aquaria.

Parthenogenesis.
To determine the general incidence of parthenogenesis, 1053 eggs collected over 105 days from a habitat with 48 females were incubated and monitored for embryo development. Five eggs from Egg cases classified as fertile were assigned to categories: hatched, parthenote or embryonic death. Egg cases classified as infertile had ova without evidence of embryonic development. Empty or wind egg cases did not contain ova.  14) were genotyped and all hatchlings and seven of the nine embryos were parthenotes, which accounts for 1.1% of eggs laid. If the remaining embryos that could not be genotyped (n = 14) were all parthenotes, the maximum incidence of parthenogenesis in this population would be 2.5%. The minimum number of females responsible for parthenotes was four accounting for 8.3% of the 48 females, however some females in the population may not have been laying eggs, making the incidence higher ( Table 4). Two of 20 inseminated females laid eggs with embryos that developed via automictic parthenogenesis, confirmed by homozygosity at all microsatellite loci. The first female that reproduced via parthenogenesis (ID 854) laid a pair of eggs four days after AI, one was infertile and the other was a parthenote. The earliest fertile eggs from AI including all females were laid 6 days after AI. Following the production of this parthenote, an additional 14 eggs were laid and seven hatched and were confirmed to be the result of AI (Fig. 4). There were 10 days between laying of the parthenote and the next AI fertilized egg. The second female that reproduced via parthenogenesis (ID 307) laid 28 eggs but only three hatched. The first egg that hatched was one of a pair of eggs laid 13 days after insemination, preceded by a pair of infertile eggs laid seven days after fertilization. This hatchling was confirmed fertilized from AI. An additional 20 infertile eggs were laid during the next 74 days followed by the last eggs laid on days 87 and 91 after AI. One egg of each of the final two pairs was infertile but the other developed via automictic parthenogenesis and hatched. There were 80 and 83 days between the AI fertilized egg and each parthenote respectively. The incidence of parthenogenesis (parthenote/total number of eggs) was 0.71% among this group of females, compared to 1.1% described above.

Discussion
This study demonstrates that offspring can be produced from AI of whitespotted bamboo sharks with sperm coldstored for up to 48 h. Following a single insemination, fertilized eggs were obtained for a period of 1-107 days up to 121 days after AI, confirming sperm storage for this species. Proof of paternity for whitespotted bamboo shark hatchlings was confirmed using existing and newly developed microsatellite loci, but also identified parthenotes within the same clutch of fertilized eggs for two females. Whitespotted bamboo shark semen was able to be collected throughout the year. In the wild, these sharks reproduce seasonally with documented seasonal changes in the gonadosomatic and hepatosomatic index for both male and females 3 . However, semen collection was not attempted in these studies, so it is possible that while the gonadosomatic index of males changed seasonally, semen may have been present year-round. Additionally, there may be changes in semen quality that accompany a seasonal reproductive pattern, as has been described for seasonal terrestrial species 36 but which were not examined in this study. www.nature.com/scientificreports/ Whitespotted bamboo shark semen was confirmed to include spermatozeugmata, sperm aggregates with heads aligned and embedded in a binding matrix 21 . Spermatozeugmata are observed for many shark species 21,37 and may serve to minimize sperm loss during copulation, increase efficiency of male and female sperm storage, or preserve longevity and motility of sperm during storage 21 . In this study a single insemination produced fertilized eggs for up to 121 days, demonstrating sperm storage for nearly 4 months. This duration of sperm storage and oviposition is similar to females newly collected from the wild in the months prior to the egg laying season and maintained in aquaria 3 as well as females that have been maintained in aquaria for several years 4,6,12 , demonstrating the safety and efficacy of AI to facilitate reproduction of this species under managed care.
Progressively motile spermatozoa had persistent flagellar motion and rotated or rolled about their long axis as has been described for other sharks 16,22 and rays 19 as well as other vertebrates 38 suggesting both motions are necessary for movement within the female reproductive tract and fertilization. Occasionally, individual progressively motile spermatozoa demonstrated periodic bursts of increased rapid forward progression especially when encountering cellular debris blocking their path forward. Spermatozoa heads were not flexible and therefore the increased forward speed may assist penetration of the protective layers surrounding the egg and be beneficial for fertilization. Spermatozoa also displayed reverse motility on occasion, similar to what has been described for other shark species 16,22 . Reverse motility is not a common spermatozoa characteristic and the role in sharks is not yet clear but could be a mechanism to liberate sperm from spermatozeugmata or be related to sperm binding sites during storage in the nidamental gland 22 . Further studies are needed to understand the significance and role of reverse motility for shark spermatozoa.
Removing the seminal plasma and resuspending in ASW resulted in greater sperm longevity of cold-stored sperm than cold storage of raw semen alone and sperm membrane integrity was maintained longer than sperm motility. These findings are in keeping with observations in the freshwater ocellate stingray Potamotrygon motoro, where extension of semen in saline or HBSS resulted in longer durations of motility than raw semen alone 14 . Loss of motility before PMI was observed, which was similarly observed in teleosts 39,40 . Insemination of females with raw semen cold-stored raw for 24 h or 48 h resulted in offspring production, but the low fertilization rate (7%) using 48 h semen suggests that washing sperm free of seminal plasma prior to shipping may confer some benefit to sperm survival prior to AI and should be investigated in the future.
The role of the male siphon sac in insemination supposed from mating observations is to aid in transfer of semen to the female through the clasper, presumably admixing the ejaculate with seawater, which would induce motility in the sperm. To attempt to investigate this role, ejaculates were either inseminated raw or extended 1:1 in ASW to mimic the mating process and induce sperm motility. Both inseminate preparation techniques produced a similar number of fertile eggs, suggesting uterine fluids and mixing of semen with seawater in the female's cloaca during natural mating may activate sperm motility 16 . Significant variation in reproductive success was observed amongst females, with some females laying few fertile eggs within a short period of time after insemination, while others produced fertilized eggs over nearly Table 4. Whitespotted bamboo shark Chiloscyllium plagiosum female and parthenote microsatellite loci. *P1-P12 eggs were laid by females housed without males and not inseminated, AI-P eggs were laid by inseminated females; -Indicates failed loci; missing data represent loci that could not be determined by genotype reconstruction. www.nature.com/scientificreports/ four months. Females that produced fertile eggs also laid unfertilized eggs near the end of their laying period suggesting that they may not have had sufficient viable sperm stored to fertilize all ova. This pattern could be due to improper timing of insemination. For females where AI of both uteri was bilaterally successful, the number of fertilized eggs cases laid was similar to that of naturally mated, recently collected in situ females 3 and naturally mated aquarium females 4,5 . Of the 20 females inseminated, five did not produce any fertile eggs following insemination, including two females that were inseminated with the lowest number of sperm, possibly indicative of insufficient sperm numbers. The remaining three females received similar concentrations of sperm to fertile females inseminated with semen from the same male, though using different ejaculates, which likely precludes sperm placement issues, and might highlight gamete incompatibility and individual female fertility differences. A wide range of fertility rates sometimes was observed for the same inseminate. This suggests that latent factors associated with the female may be more important in determining the success of a trial than inseminate characteristics, which has been shown for other species 24,41 . Females sometimes suspended egg laying for weeks or months after insemination. Similar to the unpredictable laying habits that were observed for this study, large deviations in the number of eggs laid by females in consecutive reproductive seasons has been reported for whitespotted 5 and brownbanded bamboo females 42 . Timing of insemination trials is another important factor that may affect success of AI procedures. Actively ovipositing females were chosen for AI trials but in situ females mate three months before their laying season 3 . This temporal separation of a seasonal cycle for in situ sharks may be part of the reason for the observed unpredictable oviposition after AI for sharks in managed care. Zoos and aquariums manage closed populations by recommending pairings of animals that sometime require transport of individuals between institutions so the development of assisted reproductive techniques, including AI, to manage breeding is one solution that avoids risks associated with animal transfers.

Female ID
An important consideration for AI studies in species that have facultative parthenogenesis as a reproductive strategy, is confirming parentage 7 . During this study, six females or 12% of the population laid one or more parthenotes during the study, two females confirmed from AI trials and minimally 4 additional females from monitoring the general population. Importantly, three parthenotes among fertilized hatchlings were confirmed in the same clutch for two females, underscoring the ability to switch reproductive modes within a very short timeframe, 10, 80 and 83 days, and in the same reproductive season. Previously the minimum time for switching between parthenogenesis and sexual development was nearly a year and between consecutive reproductive cycles for the same eagle ray female 34 . To our knowledge, this is the first occurrence of both pathenogenesis and sexual reproduction occurring within the same reproductive cycle for any fish.
Parthenogenesis is observed most often among females maintained without a male in managed care, or who are assumed to have been unsuccessful in finding a mate in the wild (sawfish). However, facultative parthenogenesis may be more prevalent than realized and occur normally in a small percentage of offspring for many shark and ray species. The realization that parthenotes may occur within the same clutch as sexually produced young, raises important questions for population managers. It should be noted that in this study females were housed without a male, and it is possible that there may be environmental drivers to initiate parthenogenesis that are not impacted by AI. Similar low percentages in background rates of parthenotes of females maintained without a male lend credence to this hypothesis. Future studies should include measuring the degree of incidence of parthenotes in populations with and without a male, to generate data that will allow more accurate population modeling. Microsatellite markers are commonly used to distinguish between individuals, establish relatedness, and prove paternity or parthenogenesis for whitespotted bamboo sharks; but because of a lack of diversity in the existing markers, possibly due to repeated pairings of the same individuals within an institution that leads to inbreeding, additional loci were required to confirm parentage 7,43 .
Results demonstrate AI combined with cold storage of semen is a practical tool for gene transfer between institutions or populations. Future research should include developing protocols for cryopreservation of shark sperm to increase the flexibility for timing of AI procedures and preserve gametes for transfer between populations worldwide. This provides a mechanism for moving genes without the need to transport sharks, removing a welfare consideration of potential transport stress. Additionally, the long-term use of this technique might be to augment in situ populations, allowing gene flow into, and out of wild populations, utilizing semen collected under short-term restraint as a sustainable resource, while eliminating concerns of long-term international transport that permanently depletes the same genes. www.nature.com/scientificreports/ through the urogenital papilla and slightly inclined left or right into an ampulla. Semen was extracted using gentle suction from a 1 ml or 3 ml syringe. For manual semen expression, bamboo sharks were restrained upright and out of the water, excess seawater drained from the cloaca and semen expressed using moderate bilateral ventral pressure on the region of the body immediately cranial to the pelvic girdle 11 . Semen was collected into sterile 1.5 ml conical microcentrifuge tubes and kept at ambient (~ 24 °C) temperature until processing. Semen samples were collected from individuals staggered throughout the year to encompass all months except September and December.

Methods
Semen and sperm assessment. Semen volume was measured using a calibrated pipette. Semen osmolarity was measured using a freezing point depression method (Fiske Model 210 Micro-Osmometer, Advanced Instruments Inc., MA, USA) and pH using pH strips (ColorpHast, EMD Millipore, Billerica, MA, USA). Sperm concentration was determined using a hemocytometer after diluting semen with fresh water (1:1000) to render the sperm immotile. Motility was assessed using 3 μl raw or diluted semen (1:100 artificial seawater (ASW), 1050 mOsm, Sigma S9883, St Louis, Mo. USA). Total motility (%) was defined as the number of moving sperm and progressive motility (%) was assessed as any sperm moving with forward progression. Status was assessed on a 0-5 scale with 0 for no movement and 5 for very rapid linear progression 45 . Sperm morphology was assessed using phase contrast microscopy by examining one hundred sperm per ejaculate at 100 × oil immersion and reported as a percentage. Sperm plasma membrane integrity (PMI), previously validated for use with elasmobranchs 22 , was assessed by incubating ASW diluted sperm suspensions in the dark for 10 min with 200 nM SYBR-14 and 24 µM propidium iodide (LIVE/DEAD Sperm Viability Kit L-7011, Molecular Probes, Inc., Eugene, Oregon, USA) and counting 100 cells using an epifluorescence microscope (Olympus B-Max 60) with filter cube U-M51005 for dual wavelength excitation. Spermatozoa fluorescing green over the head region were assessed as plasma membrane intact, and sperm fluorescing partially red or red over the head region were assessed as plasma membrane damaged 46 .
Acrosome presence was investigated using fluorescein isothiocyanate conjugated Arachis hypogaea agglutinin (PNA-FITC). Semen smears were made using 20 μl of semen diluted 1:100 with ASW evenly spread onto an alcohol-cleaned glass slide and allowed to air-dry. After drying, the slide was flooded with 200 μl PNA-FITC (100 μg/ml) and incubated in a dark humidified incubation chamber for 15 min at room temperature. The smear was rinsed to remove excess stain and a wet-mount examined using an epifluorescence microscope (Olympus B-Max 60) with filter cube U-M51005.
Confocal and scanning electron microscopy. An aliquot of raw semen was preserved 1:10 in 0.1 M Sorenson's phosphate buffer supplemented with 0.02% CaCl 2 , 0.35 M sucrose, 3.2% paraformaldehyde, and 2.5% glutaraldehyde for microscopy. Preserved sperm was stained for PMI, as described above, and imaged using a laser-scanning, confocal microscope (Zeiss 710, Thornwood, NY, USA) coupled with a Zeiss Axiophot inverted microscope in line scanning mode with a C-Apochromat 40 × Korr M27 (NA 1.2) water immersion objective, and ZEN software (Zeiss, Jena, Germany). Imaging was performed using differential interference contrast and fluorescence to highlight the boundary between the head and midpiece. Fluorescence was accomplished with laser excitation at 488 nm and 561 nm and emission collected between 500-550 nm and 575-610 nm. The length of the acrosome, head, midpiece and flagellum for 10 sperm per shark from five sharks were measured from digital photomicrographs using Fiji 47 . For scanning electron microscopy, preserved semen was dehydrated in increasing concentrations of ethanol and critical point dried (Tousimis Autosamdri-815B, Rockville, MD). Samples were mounted onto aluminum stubs, sputter coated with Au-Pd (Leica EM ACE600, Wetzlar, Germany) and examined using a Hitachi S-4700 scanning electron microscope (SEM).
Short-term semen storage. Cold storage of whitespotted bamboo shark sperm was investigated using fresh ejaculates from four sharks with total motility of ≥ 75%. Two medias were investigated: artificial seawater (ASW), a simple osmotically balanced salt solution and Hank's balanced salt solution (E-HBSS; 5 mM CaCl 2 , 3 mM MgCl 2 , 6 mM KCl, 0.281 M NaCl, 1 mM NaH 2 PO 4 , 8 mM NaHCO 3 , 6 mM glucose, 0.1 mM trimethylamine N-oxide (TMAO), 0.35 M urea, 0.5 mM Na 2 SO 4 , pH 7), an osmotically balanced salt solution with an energy substrate (glucose), to determine if a supplemented medium would be more supportive of cold-stored sperm longevity. Ejaculates were divided into three aliquots and stored raw or extended 1:5 in ASW or after washing to remove seminal plasma (diluted 1:10, centrifuged at 800 × g for 3 min and the sperm pellet resuspended 1:5 in ASW or E-HBSS). Aliquots were stored at 4 °C and evaluated as described previously for total motility and PMI on days 1, 2, 3, 5, 7, 12 and 18.
Artificial insemination. Whitespotted bamboo sharks (n = 20) housed without a male for 2 years were selected for artificial insemination (AI). Females were housed individually in 500-1500 L habitats to monitor egg laying and embryo development six weeks before insemination. In preparation for AI, females were anesthetized using Propofol 2.5 mg/kg intraveneously or MS-222 50-75 mg/L immersion. Once anesthetized, the female was held in dorsal recumbency with her head and gills submerged and semen transferred using a sterile, semi-rigid 14 cm 3.5 Fr polypropylene catheter inserted 8-10 cm through the cloaca and diverted laterally into an oviduct, delivering half of the inseminate to each side of the paired reproductive tract. If an egg case was present in the oviduct, semen was placed cranial to the egg case by gently manipulating the catheter around the egg case or manually extracting the egg case before insemination.
Habitats were checked daily for egg cases and hatchlings. Egg cases were tagged with the date of oviposition. Egg cases that did not contain a yolk (referred to as 'wind cases' 48 ) were discarded and not included in analyses. Eggs were candled to confirm fertility by observation of a moving embryo. Egg cases collected before AI also www.nature.com/scientificreports/ were monitored for potential embryo development but were not included in data analyses. Oviposition was monitored for 70-216 days after insemination. Females were inseminated with freshly collected or cold-stored semen that was raw or diluted 1:1 with seawater, at concentrations of 3-750 × 10 6 total sperm (Supplemental Table 1). Cold-stored semen transferred between institutions was shipped overnight at 4 °C in a stryofoam box, containing a cold pack insulated from direct contact with the semen sample. For AI procedures, samples were equilibrated to room temperature prior to insemination. Motility for all inseminates was ≥ 90% with the exception of one 24-h and one 48-h cold-stored sample (Supplemental Table 1). Fertility was calculated as the percentage of eggs that developed an embryo, and hatchability as the percentage of eggs that hatched out of the total number of egg cases laid. Hatchability and fertility were adjusted to reflect hatchlings derived only from AI by excluding confirmed parthenotes (see below). Fertility duration was calculated as the number of days each female laid fertile eggs.

Parthenogenesis.
To determine the general incidence of parthenogenesis in whitespotted bamboo sharks, all eggs laid during 105 consecutive days were collected from a habitat housing 48 mature females but no males. Eggs were incubated and monitored for development as described previously. The frequency of parthenogenesis was calculated as the ratio of the number of confirmed parthenotes (described below) to the total number of eggs (excluding wind cases) and the total number of fertile eggs. Microsatellite data from the embryos and hatchlings was used to estimate the minimum number of parthenote-laying females by reconstructing maternal genotypes. For one sample (P8) the offspring was consistent with a parthenote and it was included in analyses as a parthenote, however, missing loci prevented confirmation.

Paternity.
Whole blood or a fin clip from the trailing edge of the first or second dorsal fin was collected from dams, sires, embryos and hatchlings and stored in DMSO buffer (20% DMSO, 250 mM EDTA, saturated NaCl, pH 7.5). Paternity following AI was confirmed using techniques and four microsatellite loci previously developed for this species 7,43 , together with five new species-specific microsatellite loci developed for this study (Cpl930, Cpl962, Cpl1141, Cpl1161, and Cpl1163) following existing methodology 49 51 were combined with 8.5 µl HiDi Formamide and run on an ABI 3730 DNA Analyzer (Thermo Fisher Scientific). Genotypes were scored using Geneious v. v.10.0.3 (http:// www. genei ous. com) 52 . Paternity was determined by matching microsatellite alleles between the putative sire and dam to 88 of 112 developing embryos or hatchlings. Homozygosity for maternal alleles at all microsatellite loci coupled with a lack of paternal alleles was used to identify parthenotes 7 .
Descriptive statistics were calculated for semen and sperm parameters. Changes in PMI and motility during cold storage were examined with linear mixed effects models fit by restricted maximum likelihood with day, storage method (raw, E-HBSS or ASW) and their interaction modeled sequentially as fixed factors and repeated measures collected from the same fish accounted for by including shark as a random factor in models. Model selection was based on significance of a likelihood ratio tests between models fit by maximum likelihood that differed in fixed effects only. Model parsimony was maximized using Bayesian Information Criterion. Model residuals were examined using plots for normality, non-linearity, homoscedasticity and outliers. Pearson correlation was used to determine the relationship between motility and PMI for cold-stored semen samples. Differences in fertility for AI trials between ejaculates split and inseminated raw and diluted were evaluated using a paired t-test. Statistical analyses were conducted using R Version 4.0.0 53 with ggplot2 Version 3.3.0 54 , lme4 Version 1.1-23 and a critical probability level of 0.05.

Data availability
The datasets generated during and/or analysed during the current study are included in this published article (and its Supplementary Information files) or available from the corresponding author on reasonable request. Sequences for new whitespotted bamboo microsatellite loci have been deposited into Genbank, Accession numbers MT237441-MT237445.