|Home | About | Journals | Submit | Contact Us | Français|
In reptiles, the mode of reproduction is typically sexual. However, facultative parthenogenesis occurs in some Squamata, such as Komodo dragon (Varanus komodoensis) and Burmese python (Python bivittatus). Here, we report facultative parthenogenesis in the green anaconda (Eunectes murinus). We found two fully developed female neonates and 17 undeveloped eggs in the oviduct of a female anaconda isolated from other individuals for eight years and two months at Ueno Zoo, Japan. To clarify the zygosity of the neonates, we analyzed 18 microsatellite markers of which 16 were informative. We observed only maternal alleles and no paternal alleles for all 16 markers. To examine the possibility of the long-term sperm storage, we estimated allele frequencies in a putative parental stock by genotyping five unrelated founders. If all founders, including the mother, are originated from a single Mendelian population, then the probability that the neonates were produced by sexual reproduction with an unrelated male via long-term sperm storage was infinitesimally small (2.31E-32 per clutch). We also examined samples from two additional offspring that the mother delivered eight years before her death. We consistently observed paternal alleles in these elder offspring, indicating that the mother had switched from sexual reproduction to asexual reproduction during the eight years of isolation. This is the first case of parthenogenesis in Eunectes to be validated by DNA analysis, and suggests that facultative parthenogenesis is widespread in the Boidae.
Among vertebrates, parthenogenesis, the occurrence of unisexual lines whereby reproduction occurs without any involvement of males or their sperm, is most common amongst reptiles. Obligate parthenogenesis (OP) has been reported in more than 20 species of lizards and one species of snake (Indotyphlops braminus) . Facultative parthenogenesis (FP), occasional occurrence of parthenogenesis, in individuals of a species that normally reproduce sexually occurs in at least five families—Boidae, Pythonidae, Viperidae, Acrochordidae and Colubridae .
Two modes of FP have been proposed. Systematic facultative parthenogenesis of some species of insect and other invertebrates is thought to be highly adaptive owing to the high hatching success of unfertilized eggs by automixis . The other is accidental FP, referred as Tychoparthenogenesis , observed in some insects and vertebrates. The associated very low hatching rate suggests that this form of parthenogenesis is some form of reproduction error . However, accidental FP can be also interpreted as an emergency reproduction after the long period of isolation from mates. The low hatching rate can be explained by effects of multiple deleterious variations in the parthenogens, that are maintained as heterozygotes in sexually reproducing populations.
Two types of accidental FP occur in snakes. Type A, reported from Alethinophidia, Boidae and Pythonidae [4–8], is characterized by exclusively female offspring. Type B, reported from Caenophidia [9–15], is characterized by exclusively male offspring. Until recently, all snake species were thought to be heterogametic in females (ZW), so a number of mechanisms for the production of homogametic sons (ZZ) by Type B parthenogenesis appeared plausible. In contrast, mechanisms to produce parthenogenetic daughters by Type A parthenogenesis involved proposing WW individuals, unlikely to be viable except in rare cases. Recently, XY sex chromosomes have been discovered in Boidae and Pythonidae  suggesting that the two different parthenogenetic types in snakes simply correspond to the two different modes of sex-determination. The suggestion is that Type A occurs in species with heterogametic males (XY) to yield XX parthenogenetic females, whereas Type B occurs in species with heterogametic females (ZW) to yield ZZ parthenogenetic males.
Three species of Boidae exhibit facultative parthenogenesis that has been validated by DNA evidence, namely, Boa constrictor , Epicrates maurus  and Epicrates cenchria cenchria . Here we report a case of parthenogenetic reproduction in a long-term isolated female green anaconda Eunectes murinus, validated by DNA evidence using microsatellite markers. Although parthenogenesis has been reported before in E. murinus , our findings are the first to definitively report facultative parthenogenesis in Eunectes based on DNA analysis.
A female wild specimen of the green anaconda, the focal mother (Emu-01, Table 1) was caught in Guyana and brought to Ueno Zoo, Tokyo, Japan, on September 7, 2007. One month later, on October 16, she gave birth to 19 neonates, two of which we retained (2007OS-1 (Emu-04) and 2007OS-2 (Emu-05), Table 1). She was kept isolated from males during captivity, and on November 8, 2015, she died of pneumonia. Autopsy revealed two developed fetuses (2015OS-1 (Emu-02) and 2015OS-2 (Emu-03), Table 1) in her oviduct and 17 undeveloped eggs (Fig 1a). Both fetuses were almost fully developed but dead (Fig 1b and 1c). Examination of gonads by dissection confirmed both to be female.
Genomic DNA was extracted from muscle tissue of each of the focal mother and the four offspring (Emu-01 to Emu-05,Table 1) using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). Genomic DNA was also extracted from shed skin of additional five unrelated specimens (Emu-06 to Emu-10, Table 1) following the protocols of Fetzner (1999)  with slight modification.
We constructed Illumina NGS library for the focal mother (Emu-01, Table 1) using an Illumina TruSeq DNA PCR-Free LT Sample Prep Kit (Illumina, Tokyo, Japan). We performed 351-bp single read sequencing using MiSeq (Illumina). Filtering the raw reads using PRINSEQ  yielded 6,069,144 clean reads with a modal value of 34 for sequence quality. Using a custom perl script, we extracted 15,225 sequences harboring trinucleotide motifs with more than 10 repeat units as candidate sequences for microsatellite markers. We selected 18 of these sequences for further analyses (Genbank ID LC179548-LC179565) (Table 2).
We designed PCR primers for the candidate marker region, and added M13 sequence (5’-cacgacgttgtaaaacgac-3’) to the 5’ end of one of the primers for the fluorescent labeling of PCR products. After the initial PCR reaction, we performed second-round PCR for labeling with ROX-labeled M13 primer (5’-cacgacgttgtaaaacgac-3’). The labeled PCR products were examined by fragment analyses on an ABI3130 Sequencer using GeneMapper software version 4.1 (Thermo Fisher Scientific). After the examination of the 18 candidate markers, we could optimize genotyping condition for 16 markers. We genotyped the all ten specimens.
We assumed that the unrelated founders including the focal mother (Emu-01, Emu-06 to Emu-10) are from a Mendelian population in Hardy-Weinberg equilibrium. Dividing the observed number of alleles by number of haplotypes (2n = 12) yielded estimates of allele frequencies for each marker (S1 Table). Under our null hypothesis, the genotype of a putative father of the two neonates is drawn from this allelic distribution.
Then our observed genotypes for the two neonates arise from one of the three cases where p1 and p2 are the frequencies of maternal alleles:
The joint probability is the product of Pk for the 16 makers assuming that they are independent.
We summarized the genotyping results in Table 3. We found the focal mother was heterozygous for 14 out of the 16 microsatellite markers. The two female neonates consistently had only one allele inherited from the focal mother. There were no non-maternal alleles observed in either of the neonates. The lack of paternal alleles with apparent homozygosity of maternal alleles strongly suggests that the neonates were produced by parthenogenesis by the focal mother.
The joint probability that two homozygous neonates were produced by sexual reproduction by the focal mother and the putative father, under the assumption that they were drawn from the population characterized by the five unknowns and the mother, and of linkage equilibrium for the markers, was infinitesimally small, 2.31E-32 (Table 4). We therefore excluded the possibility that the two neonates arose through sexual reproduction via long-term sperm storage and confirm our conclusion that the female neonates are parthenogenetic offspring. In contrast, paternal and maternal alleles were present in elder offspring born in 2007 with the same mother– 2007-OS1 (Emu-04) and 2007-OS2 (Emu-05). The focal mother has thus reproduced sexually at least once in the past, and has switched to parthenogenetic reproduction during the 8 years of her isolation.
Here, we report the first DNA-validated case of facultative parthenogenesis in the green anaconda, Eunectes murinus. Since facultative parthenogenesis producing only females has been reported in other boid species such as Boa constrictor , Epicrates maurus  and Epicrates cenchria , Type A facultative parthenogenesis is to date the most common form, perhaps the only form, of parthenogenesis in Boidae. Female-biased facultative parthenogenesis has also been reported in the Pythonidae–Python bivittatus , Python regius and Malayopython reticulatus –suggesting that it is the ancestral state shared in basal Alethinophidia. Recently, XY sex chromosomes have been discovered in Boa imperator and Python bivittatus , and are likely to occur broadly in the Pythonidae and Boidae, including Eunectes murinus. Female offspring produced by Type A parthenogenesis with XX or XO offspring, is now the expectation for this group of snakes subject to further investigation.
Three possible mechanisms for parthenogenesis yielding diploid offspring are recognized: central fusion, terminal fusion and gametic duplication. In species with homogametic females (XX), such as Eunectes murinus, central fusion is expected to produce female offspring only, with the same level of heterozygosity as seen in our focal mother. Since no such case has been ever reported, including the present study, central fusion is highly unlikely to be the mechanism of parthenogenesis in snakes. Terminal fusion produces female-only offspring that are largely homozygous (autozygous), but partially heterozygous owing to recombination during maternal meiosis. Gametic duplication produces female-only offspring that are completely autozygous for maternal chromosomes, even in the region where recombination occurs during maternal meiosis. Limitations of the informative markers used in previous studies of parthenogenesis in Boidae did not allow for a clear distinction between terminal fusion and gametic duplication (Booth et al 2011a using 4 informative markers ; Booth et al 2011b using 4 informative markers ; Kinney et al 2013 using 7 markers ). In the present study, the female offspring were consistently homozygous for all of 16 microsatellite markers, including 14 markers for which the focal mother was heterozygous (Table 3). The exclusive homozygosity (14/14) we observed strongly indicates that the parthenogenesis in Eunectes murinus is produced by gametic duplication, not terminal fusion. This hypothesis could be tested with high-resolution genome-wide analyses such as whole genome sequencing or reduced representational approaches such as ddRAD  or DArTSeq .
Although highly unlikely for vertebrates, there is a possibility that the parthenogenetic offspring of Eunectes murinus are haploid (XO), developed from unfertilized eggs. Such parthenogenesis by reduction of ploidy has been reported from vertebrates only in whitetip reef shark Triaenodon obesus . To exclude this possibility, it is necessary to examine karyotypes by cytological analyses or by quantitation of nuclear DNA. Unfortunately the parthenogenetic neonates described in the current study were dead upon the autopsy of their mother; we were unable to obtain live cells for cytological analyses or FACS analyses, and so cannot eliminate the slight possibility of haploid offspring.
Since the two parthenogenic offspring were female and dead, there is no direct information on their viability. However, they had no obvious morphological anomalies. A recent suspected case of parthenogenesis in Eunectes murinus has produced viable offspring, but uncertainty remains on the status of this case as no DNA evidence in support of the parthenogenesis was included . DNA-validated parthenogenetic offspring were viable in the closely related Boidae, Boa constrictor, Epicrates maurus and Epicrates cenchria cenchria [5–7]. Therefore, it is quite likely that the two female pathenogens of Eunectes murinus would have been viable had they not died.
Parthenogenetic events are noticed usually when females are held in long-term isolation, suggesting that the certain period of isolation may be a trigger for parthenogenesis. However, as parthenogenetic events have been reported also in females housed with males , the mechanism to trigger parthenogenesis remains unknown. Care must be taken to avoid the reduction of genetic heterogeneity in captive breeding programs in potentially facultative parthenogenetic species by monitoring, with DNA technologies, the occurrence of parthenogenesis.
All of the facultative parthenogenesis reported in Boidae and Pythonidae to date, including the present study, have been exclusively in captivity. Parthenogenesis has been reported in wild populations of other Alethinophidia . However, there has been insufficient effort to detect parthenogenesis in wild boid populations. Hence, parthenogenesis may well also occur in wild boid populations at a low but biologically significant rate. Facultative parthenogenesis via terminal or gametic duplication produces largely autozygous offspring and so results in genetic purging of detrimental variation from the population by enhancing natural selection . Therefore, parthenogenesis may play some role in maintenance of genetic “health” in wild population of parthenogenetic species such as Eunectes murinus. Comparison of heterozygosity in wild populations between parthenogenetic and non-parthenogenetic species may provide opportunity to test the “purging” hypothesis of deleterious alleles.
We greatly thank Suma Aqualife Park, Nihondaira Zoo, Noboribetsu Marine Park Nixe, Higashiyama Zoo and Botanical Gardens, Maruyama Zoo and Dr. Shin-ichiro Kawada, National Museum of Nature and Science for providing samples and information for their green anaconda specimen. We also thank Prof. Hidenori Tachida for his technical advice. Prof. Arthur Georges assisted greatly with English expression and grammar on a late draft of the manuscript. All of the authors declare no conflicts of interest associated with this manuscript.
The authors received no specific funding for this work.
All relevant data are within the paper.