The present study analysed the transcriptome profiles of gonadal tissues from C. gibelio using RNA-seq, specifically to identify DEGs in ovaries associated with reproduction in triploid gynogenetic females and diploid sexual females. We also analysed the transcriptome profiles of ovaries and testes in males of C. gibelio, and the two closely-related species C. auratus and C. carpio. A total of 1728 genes were significantly upregulated or downregulated in asexual females of C. gibelio compared to sexual females. The transcriptome profiles based on normalized RNAseq read counts showed a sex-dependant difference for both - all transcribed genes or reproduction-associated genes, with an overall similarity between gynogenetic and sexual females of C. gibelio and females of C. auratus, and an overall similarity between the males of the two Carassius species.
GO term overrepresentation analyses and KEGG pathway enrichment analyses indicated an overall overexpression of genes involved in meiosis and cell cycle control (cell cycle, negative regulation of nuclear division, negative regulation of cell cycle process, oocyte meiosis, and synaptonemal complex assembly), oocyte maturation (egg coat formation, structural constituent of egg coat, and calcium ion binding) and fertilization (binding of sperm to zona pellucida, positive regulation of acrosome reaction). Calcium ion binding, which plays critical roles in fertilization and early development (for review, see Whitaker (67)), was also overrepresented in sexual females. This suggests that the regulation of oogenesis, as well as the response of oocytes to sperm cell binding, differ between sexual reproduction and gynogenesis, where the eggs are only activated by the sperm cell (for review, see Schlupp (68)). An overall downregulation of meiotic and reproduction-associated genes was also reported in Poecilia formosa, a gynogenetic fish species of the Amazon basin, compared to its sexual parental ancestors, P. mexicana and P. latipinna (19). Similar results were reported in invertebrates that use cyclical parthenogenesis, such as the planktonic crustacean Daphnia, rotifers, and aphids, where the sexual forms upregulate genes involved in cell cycle control, meiosis, oogenesis, and oocyte maturation (69–72).
On the basis of ovarian transcriptome profiles, we identified around 100 reproduction-associated genes related to oocyte meiosis, oogenesis, embryogenesis, hormone signalling, and fertilization that were differently expressed between sexual and gynogenetic females; the expression pattern of a set of 17 selected genes based on the basis of RNAseq was validated by RT-qPCR. We also specifically analysed 40 meiosis-related genes inferred by previous studies (54–56, 62, 63). We showed that sexual females upregulated several meiosis-associated genes involved in recombination and crossover and in DNA double-strand break formation during meiosis, including spo11, msh2, pds5b, sbk3, stag1a, and rec114. Two components of the minichromosome complex (mcm4 and mcm9), involved in crossover inhibition during meiosis (73), as well as syce1, a component of the synaptonemal complex that forms between homologous chromosomes during recombination, were also upregulated in sexual females (74–76). Sexual females also upregulated genes involved in oocyte maturation, such as emi1 (also named fbxo5), a major F-box constituent of the E3 ubiquitin ligase protein that regulates the anaphase promoting complex (APC) during meiosis and mitosis (77–80); and spinw, a major maternal transcript expressed in oocytes during early development. The importance of spindlin in oocytes to embryo transition in C. gibelio has been established (81). Furthermore, several genes involved in cell cycle regulation, including three members of the Ras/MAPK family, hrasa, hrasb and rasa1b, which encode GTPases controlling cell growth, division, and differentiation (82–85) through the action of mitogen activated protein kinases (86), were also more expressed in sexual females. This suggests that cell cycle control regulation differs between sexual and gynogenetic females of C. gibelio.
In accordance with our results, gynogenetic P. formosa was shown to underexpress meiosis-related genes, including sbk3, setd7 and stk32c, compared to its supposed sexual ancestors (19). Similarly, in cyclically parthenogenetic Daphnia, meiosis-related genes, including genes related to the spindle assembly checkpoint, the APC, and meiosis chromosome segregation, were upregulated during sexual reproduction (71). In particular, spo11, which encodes a topoisomerase involved in chromosomal recombination during the meiotic prophase, was also described as an important player in the meiosis-to-parthenogenesis transition in pea aphid (87), although it was not reported in asexual P. formosa (19).
However, our study also revealed that meiosis pathways were not fully disrupted in gynogenetic females of C. gibelio. They retained detectable expressions of all reproduction-associated genes identified, including meiosis-specific genes, in contrast to P. formosa, where some meiosis-related genes were not expressed (19). According to our analyses, several of the core meiosis specific genes, such as dmc1, mlh1, mnd1, mre11 and genes of the msh family (55, 55, 56, 62, 63), did not show significant differences in expression between sexual and gynogenetic females of gibel carp. Gynogenetic females even upregulated rad1, a member of the cell cycle checkpoint, also involved in the recombination process during meiosis; rnf212, involved in meiotic recombination; and mad2l2, involved in the spindle assembly checkpoint; as well as meiosis-specific genes that were previously found to be downregulated in gynogenetic P. formosa, such as b4galt, clk4, dmrta2, grapb, and rasl11b (19). However, these results are in accordance with a study suggesting that meiosis is retained even in gynogenetic strains of C. gibelio in North-east Asia (88). Furthermore, meiosis genes were reported not to be necessarily associated with sexual reproduction, since asexual amoeba constitutively expressed meiosis-associated genes (54). Similar results were reported also in rotifers, where no meiosis-specific genes were differently expressed between parthenogenetic and sexual forms (70), and cyclically-parthenogenetic Daphnia, which was shown to express meiosis-specific genes during the parthenogenetic phase (89). In the pea aphid, several oogenesis and cell cycle-related genes were also upregulated during the asexual reproduction phase (69).
Our results reveal an overall upregulation of pathways related to oocyte maturation in sexual females. They upregulated buc, involved in the formation of Balbiani bodies in the oocytes and germ plasm assembly, including follicular epithelium morphogenesis (90). This gene plays a key role in the specification of oocyte anterior/posterior polarity through interactions with the RNA-binding proteins, such as rbpms2, a coactivator of transcriptional activity involved in meiosis and oogenesis (91). Sexual females of C. gibelio also upregulate genes involved in progesterone-mediated oocyte maturation, such as members of the plexin and Wnt families. The Wnt pathway regulator lbh, previously reported to be upregulated in females during oocyte maturation in C. gibelio, was also more expressed in sexual females in our study. Similarly, in aphids, genes involved in oocyte axis formation were found to be upregulated during the sexual phase (72). Furthermore, our analyses support an overall upregulation of sperm-egg recognition and fertilization pathways in sexual females. They upregulated calm3a, a member of the calmodulin family responsible for calcium-dependant signal transduction following sperm binding, as well as plcb4, a phospholipase involved in oocyte fertilization (92). In addition, sexual females upregulated components of the zona pellucida, the extracellular matrix surrounding the oocyte involved in sperm-egg recognition (93). A gene encoding a Ca2+-dependant C-type lectin, which was shown to be translocated in cortical granules during oocyte maturation and involved in sperm-egg recognition and fertilization in C. gibelio (94), was also significantly upregulated in sexual females. These findings highlight the importance of oocyte maturation, sperm-egg recognition, and fertilization pathways in the coexistence of sexual and asexual females.
Inversely, some genes involved in oocyte development, such as DAZ-like genes, were not differentially expressed between gynogenetic and sexual females of gibel carp in our study, while others, including bcl2; the oocyte specific histone h2af1o, which plays a key role in fish embryogenesis (95); and several members of the FGF family, which promote meiosis and maturation of the oocytes (96), were even more expressed in asexual females than in sexual ones. Oocyte maturation and sperm cell binding pathways are not expected to be disrupted in asexual females, since they produce oocytes. Furthermore, gynogenetic C. gibelio females still require sperm cell binding to activate the eggs (68, 97). The overexpression of some oogenesis-related genes was also reported in aphids during the parthenogenetic phase of their life cycle (69). Furthermore, the downregulation of uhrf1, an oocyte-specific epigenetic regulator (98) in sexual females of C. gibelio, also reported in aphids (69), suggests a difference in the epigenetic regulation of oogenesis between sexual and asexual forms. Hence, these results suggest that many genes and pathways are involved in both parthenogenetic oogenesis and sexual oogenesis in C. gibelio. However, gene expression differs between the two reproduction forms. It is noteworthy that members of the same gene family can be up- or downregulated, such as members of the zona pellucida and F-box families. Such divergent expression, also reported in Daphnia (71), may suggest functional divergence among members of the same multigenic families.
Our analyses also suggest differences in hormonal signalling and sex differentiation processes between sexual and gynogenetic reproduction. Components of the GnRH signalling pathway, and genes linked to ovarian fertility, such as the gene encoding the luteinizing hormone/choriogonadotropin receptor (lhcgr), were more expressed in sexual females. The TGF-β signalling pathway, involved in many physiological processes including sexual differentiation in fish (99–101), was also differently regulated between gynogenetic and sexual females of C. gibelio. Sexual females upregulated smad genes, involved in oogenesis, ovarian function, and folliculogenesis via the negative regulation of TGF-β signalling. Regarding gynogenetic females, they upregulated two dmrt genes. These genes were shown to promote male differentiation and repress female-specific differentiation of the gonads, and they are also involved in brain sexual differentiation (64, 104, 105) as well as in XY reversal in sex-alternating fish species (64). Gynogenetic females of C. gibelio also upregulated ncoa2, a transcriptional coactivator of steroid receptors and nuclear receptor, as well as sox8, involved in female sex determination (106), meiotic progression, and embryonic development (107), and inhibin alpha (inha), involved in steroid hormone biosynthesis. Ovarian aromatase or estrogen synthetase (cyp19a1a), a member of the cytochrome P450 subfamily involved in steroidogenesis (108) and female folliculogenesis and gonadal differentiation, was also upregulated in gynogenetic females of C. gibelio, as was oxtr, a gene encoding the oxytocin receptor, a component of the oxytocin signalling system that modulates reproductive behaviour. Our results also suggest that sexual females upregulated some genes associated with the steroid hormone synthesis pathway. The hydroxysteroid 17-β-dehydrogenase gene hsd17b1, which is both estrogenic (109) and androgenic (110), was more expressed in gynogenetic females. Furthermore, sexual females also upregulated the germ cell maintenance gene piwil2, a member of the Argonaute family involved in male fertility (111).
In this study, we also investigated the evolutionary history of C. gibelio. Ploidy changes shaped the evolution of cyprinids, particularly that of the Carassius auratus complex. This complex was formed by allotetraploidization (36, 112) and further polyploidization events have been reported in diverse lineages of the complex, including C. auratus and C. gibelio (36, 113). However, the evolutionary origin of C. gibelio is still in question. A study based on dmrt genes suggested a recent autopolyploidization event within the C. auratus complex that generated the triploid gynogenetic C. gibelio (35). However, an origin of C. gibelio by hybridization between C. auratus and C. carpio has also been proposed (114). Our SNP clustering, based on gonadal transcriptomes, using C. gibelio, C. auratus and C. carpio, suggests a close evolutionary relationship between sexual and gynogenetic C. gibelio, as well as a close relatedness between C. gibelio and C. auratus. This is in accordance with a study showing that two gene copies of four different Hox genes in the genome of gynogenetic C. gibelio are orthologous to the Hox genes of C. auratus and that one is orthologous to the Hox gene of C. carpio (114). That study suggested that triploid gynogenetic C. gibelio (3n = 15) resulted from interspecific hybridization between diploid C. auratus (2n = 100) and C. carpio (2n = 100), contributing with two sets and one set of chromosomes, respectively. However, the diploid form of C. gibelio was not included in that study. Other studies using mtDNA and hoxa2b gene sequences even suggested a more complex relationship between C. gibelio and C. auratus, where the monophyly of C. gibelio was not supported (115, 116). In addition, gene flow was highlighted between the two species (88, 115), suggesting that C. gibelio and C. auratus were conspecific and interfertile.
Ploidy changes often affect meiosis, and parthenogenetic species usually result from interspecific hybridization (8) with some exceptions (117). Polyploidy can lead to the formation of unreduced eggs whose cell cycle is arrested at the metaphase of meiosis II (118). This results in asexually reproducing species, where the offspring are clones of the mother. Unisexual fish reproduce through gynogenesis, where the sperm from males of the same or closely-related species is still required to activate the egg. Still, because meiosis pathways were not disrupted, a later genetic contribution from a sperm donor such as C. auratus and C. carpio cannot be excluded. Such a case of a complex evolutionary history was reported in the unisexual salamander Ambystoma. However, in this case, the haploid genome of the sperm donor replaced the nuclear genome, a phenomenon known as kleptogenesis (119, 120).
Our results suggest that all along their evolutionary history, asexual lines of C. gibelio did not lose the genetic toolkit for meiosis, and that the sexual reproduction genetic toolkit is not under relaxed selection, a condition also reported in asexual P. formosa (19) and snails (121). The re-acquisition of sexual reproduction in asexual species is very rare and very few cases have been reported. Either some gynogenetic C. gibelio females were able to secondarily regain sexual reproduction and to produce both diploid and triploid males, or a minority of sexual individuals still persisted within the already formed gynogenetic form and became more abundant later (56). In all cases, this led to the current sympatric coexistence of sexual and gynogenetic individuals (26, 122). Polyploidy in general, and triploidy in the case of gynogenetic C. gibelio could possibly compensate the deleterious effects of Muller’s ratchet or the accumulation of deleterious mutations by increasing the number of gene copies and favouring heterozygosity (54). The genomic incorporation of sperm-derived fragments from an exogenous species, which was reported in gynogenetic C. gibelio from aquaculture in China (27), can also favor genetic diversity in asexual lines. In C. gibelio, the combination of the advantages of gynogenetic reproduction, which allows for faster population growth (23, 23), and sexual reproduction, which provides higher resistance to parasites and higher immune gene variability (29), higher aerobic performance and better immunity (123), lower metabolic rate, and lower energy intake (124), might explain the coexistence of sexual and asexual forms, and the high adaptive abilities of this species and its invasiveness in European water ecosystems.