Formation of autotriploid Carassius auratus and its fertility-related genes analysis

Background Formation of triploid organism is useful in genetics and breeding. In this study, autotriploid Carassius auratus (3nRR, 3n = 150) was generated from Carassius auratus red var. (RCC, 2n = 100) (♀) and autotetraploid Carassius auratus (4nRR, 4n = 200) (♂). The female 3nRR produced haploid, diploid and triploid eggs, whereas the male 3nRR was infertile. The aim of the present study was to explore fertility of potential candidate genes of 3nRR. Results Gonadal transcriptome profiling of four groups (3 females RCC (FRCC), 3 males 4nRR (M4nRR), 3 males 3nRR (M3nRR) and 3 females 3nRR (F3nRR)) was performed using RNA-SEq. A total of 78.90 Gb of clean short reads and 24,262 differentially expressed transcripts (DETs), including 20,155 in F3nRR vs. FRCC and 4,107 in M3nRR vs. M4nRR were identified. A total of 106 enriched pathways were identified through KEGG enrichment analysis. Out of the enriched pathways, 44 and 62 signalling pathways were identified in F3nRR vs. FRCC and M3nRR vs. M4nRR, respectively. A total of 80 and 25 potential candidate genes for fertility-related in F3nRR and M3nRR were identified, respectively, through GO, KEGG analyses and the published literature. Moreover, protein-protein interaction (PPI) network construction of these fertility-associated genes were performed. Analysis of the PPI networks showed that 6 hub genes (MYC, SOX2, BMP4, GATA4, PTEN and BMP2) were involved in female fertility of F3nRR, and 2 hub genes (TP53 and FGF2) were involved in male sterility of M3nRR. Conclusions Establishment of autotriploid fish offers an ideal model to study reproductive traits of triploid fish. RNA-Seq data revealed 6 genes, namely, MYC, SOX2, BMP4, GATA4, PTEN and BMP2, involved in the female fertility of the F3nRR. Moreover, 2 genes, namely, TP53 and FGF2, were related to the male sterility of the M3nRR. These findings provide information on reproduction and breeding in triploid fish. Supplementary Information The online version contains supplementary material available at 10.1186/s12864-021-07753-5.


Background
Polyploid organisms have three or more chromosome sets. Triploidy, an example of polyploids, plays a vital role in the process of biological evolution and can be divided into autotriploidy and allotriploidy [1]. Allotriploids have three chromosome sets from two or more different species, whereas autotriploids have three chromosome sets derived from a single taxon.
Development of gonads is critical to fertility in sexually reproducing organisms especially in triploids and is tightly regulated by complex processes [2]. Sex determination, sexual differentiation and gametogenesis are important processes during gonadal development. Any abnormality in these events can result in infertility. Several genes implicated in sexual determination and differentiation have been reported [3][4][5]. Gametogenesis, including oogenesis and spermatogenesis, are also regulated by complex mechanisms and several regulatory genes [6,7]. A previous study explored regulation of early stages of oogenesis [8]. In addition, studies explored biological mechanisms that occur mid-oogenesis [9], and regulation of late oogenesis [10]. Spermatogenesis is divided into three steps: spermatogonial mitotic proliferation, two times of meiosis, and post-meiotic differentiation [11]. Previous studies explored the functional mechanisms of spermatocytogenesis [12], meiosis during spermatogenesis [13], and spermiogenesis [14].
Fertility of polyploids has important implications in fisheries and sustainable aquaculture. Artificial triploids of species such as Atlantic salmon, Oncorhynchus mykiss, Salmo trutta and Salvelinus fontinalis have been widely used in fish farming industry [15]. There has been a believe that triploidization causes infertility in fish. In addition, a previous study reports that the triploid can channel the energy required for gonad maturation to somatic growth, causing rapid growth rates compared with their diploid counterparts [16]. However, a different study reported that triploid fish can produce normal gametes [17]. In our previous study, Carassius auratus red var. (RCC) (female) and autotetraploid Carassius auratus (4nRR) (male) were artificially hybridized to produce hybrid autotriploid Carassius auratus (3nRR). After hybridization, the male 3nRR did not produce normal sperm, whereas the females generated dynamic eggs [18]. Analysis of meiosis-related gene expression showed that Dmc1 and Ph1 had higher expression level in female 3nRR compared with levels in the males, indicating that these genes are involved in regulating fertility of 3nRR [19]. Molecular mechanisms involving 3nRR fish in controlling fertility have not been explored fully.
RNA-Seq technology is utilised for analysis of the structure and function of genes at the organismal level, and for exploring a series of biological pathways [20]. RNA-Seq technique has been successfully used in studies various fishes in the past decade. In spotted scat species (Scatophagus argus), several candidate genes involved in reproduction and gonadal development were obtained by RNA-Seq [21]. Studies on Takifugu rubripes reported that sex-related genes play an important role at early sex differentiation stage [22]. Gonadal transcriptome profiling of triploid hybrid loaches (Misgurnus anguillicaudatus) and their diploid and tetraploid parents showed key genes implicated in low hybrid triploid fertility [17]. A study on Thunnus maccoyii reported sex and gonad-development-related genes in the gonads of Southern bluefin tuna through RNA-Seq [23]. In addition, RNA-Seq has been successfully used to analyze sex determination and differentiation related genes in tilapia [24]. In the present study, we successfully obtained triploid fish (3nRR) by crossing female RCC and male 4nRR. The diploid (2nF 1 ), triploid (3nF 1 ) and tetraploid (4nF 1 ) hybrids were then generated by hybridization of female 3nRR and male RCC. In this study, we explored important biological traits and systematically compared gonadal transcriptome of the triploid fish (3nRR) with their parents. Further, the molecular mechanism of the low fertility of the autotriploid fish was explored. The findings of this study provide information on the biological characteristics of 3nRR and mechanisms associated with fertility regulation in triploid fish.

Results
Fertility of autotriploid Carassius auratus 3nRR were generated by crossing female RCC and male 4nRR during the breeding season (Figs. 1a, b and c, 2 and 3a, b and c; Table 1). Testes of RCC and 4nRR (Fig. 4a, b) contained spermatogonia (SG), spermatocytes (SC) and a large number of mature spermatid (ST), whereas the mature sperm was not observed in 3nRR (Fig. 4c). Ovaries of RCC, 4nRR and 3nRR contained second, third and fourth phase oocytes (Fig. 4d, e, f). These results indicated that all ovaries, and RCC and 4nRR testes were fertile whereas 3nRR testes were sterile.
Eggs and water-like semen were collected during the reproductive season from two years old males and females of 3nRR, respectively (Fig. 5). Ploidy levels of the offspring resulting from a cross of female 3nRR and male RCC (Fig. 1c, d, e, f, g) were determined by measuring the chromosome number (Fig. 3d, e, f; Table 1). These analyses showed that female 3nRR produced different sizes of eggs.

Transcriptome sequencing and sequence alignment
Optical density (OD) ratio A260/A280 and RNA integrity numbers (RINs) of the RNA in 12 samples (Table 2) were 2.1 and 8.0-8.8, respectively (Additional file 1). These results indicate that all samples were free from contamination and their quality met the requirements for transcriptome sequencing.
RNA-seq from gonadal tissue samples of autotriploid fish and their parents was performed by Illumina. RNA-Seq results are presented in Tables 3 and 4. Number of clean reads from the 12 RNA-seq libraries ranged from 39,624,312 to 50,588,484. All clean reads were then aligned to the RCC genome sequences using HISAT2 software. Mapped genome reads ranged from 24,237,536 to 42,474,296, genome map rates ranged from 59.79 to 91.55 %, and unique match rates ranged from 57.84 to 85.93 %.
Hub genes related to the fertility in 3nRR were identified Eighty genes out of the DETs identified in the F3nRR vs. FRCC group related to female fertility were identified by  M4nRR group are implicated in male sterility (Additional file 7). To further identify hub genes associated with 3nRR fertility, PPI of the fertility-related genes was constructed using STRING tool and analysis was carried out using Cytoscape software. After analysis of PPI network of female fertility-related genes, 6 genes with the interaction degrees more than 15 were screened as hub genes (Fig. 9a, Additional file 8). Furthermore, PPI of male sterility-related genes showed that 2 hub genes, with degrees more than 5 showed strong interaction with other node proteins ( Fig. 9b, Additional file 9).

RT-qPCR verification
To verify RNA-Seq results, twenty-eight DETs were chosen for validation by RT-qPCR. Among the 28 DETs, 6 DETs and 7 DETs were up-regulated in the F3nRR vs.
FRCC and M3nRR vs. M4nRR groups, respectively; whereas 10 DETs and 5 DETs were down-regulated in the F3nRR vs. FRCC and M3nRR vs. M4nRR groups, respectively ( Fig. 10). Expression profiles of the twenty DEGs obtained by RT-qPCR and RNA-Seq were similar, implying that RNA-Seq results were reliable.

Discussion
Triploid animals are usually sterile and cannot form triploid populations. However, previous studies have been reporting contradicting results. Xiao et al. [25] reported that triploid Carassius auratus in Dongting water system produces normal gametes. Hu et al. [26] reported that female autotriploid hybrids (3nAUT) generated by crossing females of Carassius auratus red var. with males of autotetraploid fish produced mature eggs. However, male 3nAUT showed abnormal gonadal development and could not produce mature sperm. In the present  study, 3nRR was generated by crossing female RCC and male 4nRR. Three different ploidy offspring were then obtained by hybridization of female 3nRR and male RCC. The female 3nRR offspring were fertile, whereas male 3nRR were sterile. However, only a few studies have explored the molecular mechanisms modulating fertility of the autotriploid of Carassius auratus. In this study transcriptome analysis was used to explore the molecular mechanisms associated with poor fertility in 3nRR. Eight fertility-related hub genes of 3nRR were identified through GO and KEGG enrichment analyses, and previous published literature.
Candidate hub genes related to male sterility of 3nRR were identified Hub genes identified in the M3nRR vs. M4nRR group included several genes involved in male sterility, such as the tumor protein p53 (TP53) and fibroblast growth factor 2 (FGF2).
TP53, also known as P53 is a transcriptional regulator and tumor suppressor implicated in spermatogenesis [27]. In vertebrates, partial or complete impairment of P53 expression causes disordered meiotic divisions, which in turn causes spermatogenesis defects [28,29]. P53 mRNA and protein levels are downregulated in the testis of P53 promoterchloramphenicol acetyltransferase (CAT)-harboring mice, indicating its important role in development of spermatocytes [30]. In addition, TP53 codon 72 polymorphism in mice is involved in meiosis, implying that it plays a critical role in spermatogenesis [31]. In human, P53 gene polymorphism is higher in infertile men compared with fertile men, implying that it may affect germ cell apoptosis and increase risk of male infertility [29,32]. In the current study, analysis of expression levels of the TP53 gene showed significantly different expression levels between M3nRR and M4nRR. This finding implies that TP53 may disrupt meiosis during spermatogenesis in the male 3nRR fish causing sterility.
Fibroblast growth factor 2 (FGF2) plays essential functions in regulation of spermatogenesis and sperm physiology [33]. A study using a human model reported presence of FGF2 and FGFRs in testis and sperm, which are related with human spermatogenesis and sperm motility [34]. Furthermore, incubation of human sperm with recombinant FGF2 (rFGF2) causes an increase in number of motile cells, implying that the gene is involved in sperm motility [35]. In mouse, knock out of FGF2 induces impaired sperm production and is associated with alterations in sperm morphology and function [36]. In this study, FGF2 was significantly upregulated in M3nRR vs. 4nRR. High expression levels of the gene can cause abnormal shaping of the normal sperm, which resulted in male 3nRR sterility.  Candidate hub genes related to female fertility of 3nRR were identified Six hub genes associated with female fertility were identified in the F3nRR vs. F4nRR group including MYC proto-oncogene, bHLH transcription factor (MYC), SRY-box transcription factor 2 (SOX2), bone morphogenetic protein 4 (BMP4), GATA binding protein 4 (GATA4), phosphatase and tensin homolog (PTEN) and bone morphogenetic protein 2 (BMP2). MYC gene encodes the MYC transcription factor which is involved in cell proliferation and gametogenesis [37]. In Xenopus, C-MYC was detected in oocytes, indicating that it plays a role in oogenesis [38]. In Drosophila, MYC was involved in distribution of CTPsyn in follicle cells, implying that it plays a role in synthesizing nutrients for the developing oocytes [39]. In addition, a previous study reported that MYC plays an important role in regulation of mitochondrial biogenesis in Drosophila ovary, indicating that the gene is involved in oocyte development [40]. In Larimichthys crocea, Lc-cMYC had different expression patterns in oocytes at various stages of development, implying that it plays an essential role in oogenesis [41]. In the current study, 3nRR ovary showed low expression level of MYC which may inhibited formation of oogenesis defects during triploidization, resulting in production mature eggs.
SOX is an ancient gene family involved in oogenesis [42]. Sox genes have been explored in many organisms and can be classified into ten subgroups (A-J) [43]. In Agasicles hygrophila, AhDichaete and AhSox3 expression levels are significantly high in ovary, indicating that it plays a vital regulatory role in during ovarian development and oogenesis [44]. In Misgurnus anguillicaudatus, MaSOX3 is abundant in primary oocytes and previtellogenic oocyte cells, indicating that MaSox3 gene is involved in ovarian development [45]. In Paramisgurnus dabryanus, SOX4 was detected in the ovary, showing that it plays an important role during ovarian development [46]. In mouse, expression of SOX2 is required for establishment and maintenance of the oocyte cell [47]. In our study, SOX2 was identified in F3nRR vs. FRCC, with lower expression level in F3nRR gonad compared Bone morphogenetic proteins (BMPs) are belonging to the transforming growth factor-β superfamily of proteins, and they appear to be highly conserved [48]. A previous study reports that BMPs play a role in regulation of ovarian follicular development [49]. BMP1, BMP6 and BMP15 are implicated in ovarian development [50][51][52]. In Xenopus laevis, BMP2 gene is highly expressed during oogenesis, implying that it is an important factor in ovarian development [53]. In mouse, BMP4 regulates the number of oocytes, suggesting its role in the process of oogenesis [54]. Roles of BMP2 and BMP4 as important factors in survival and development of bovine secondary follicles were recently reported [55]. In this study, BMP2 and BMP4 genes were differentially expressed between F3nRR and FRCC, indicating that they may be involved in ovarian development.
GATA4, a member of the GATA-binding family, is highly expressed in ovarian granulosa cells [56,57]. A previous study reported that the gene is involved in regulation of ovarian development [58]. GATA4 and GATA6 knockout female mice exhibited infertility due to disrupted formation of ovaries [59]. GATA4 deletion resulted in a sterile female mice phenotype attributed to drastic reduction in number of developing follicles [60].
The mRNA for GATA4 has been reported in human ovary implying that GATA4 plays a role in ovarian folliculogenesis [57]. GATA4 was identified through transcriptome analysis in this study. GATA4 was differentially expressed in F3nRR vs. FRCC, indicating that plays important roles in ovarian development of 3nRR.
Phosphatase and tensin homolog (PTEN) protein has phosphatase activity and belongs to protein-tyrosine phosphatase superfamily [61]. PTEN is a negative regulator of PI3K-Akt signaling pathway which is involved in growth of eggs [62]. Deletion of PTEN from oocytes affects mouse fertility by interrupting oocyte growth [63]. PTEN signaling pathway associated with ovarian follicle development has been reported in human [64]. In Crassostrea gigas, PTEN is involved in insulin pathway in gonads and plays a critical role in reproduction [65]. In Drosophila, loss of PTEN is related to IIS/mTORC1 signalling, which is important for oogenesis [66]. In the present study, F3nRR showed significantly low expression levels of PTEN, which may have caused F3nRR normal fertility through regulation of oogenesis.

Conclusions
The autotriploid Carassius auratus (3nRR, 3n = 150) is generated from Carassius auratus red var. (RCC, 2n = 100) (♀) and autotetraploid Carassius auratus (4nRR,  M3nRR vs. M4nRR. The abscissa represents P value, which decreases with increase in significance in enrichment level of differentially expressed transcripts in the pathway. The ordinate represents log10 (Q value), which increases with increase in significance of differentially expressed transcripts in the pathway 4n = 200) (♂), of which the female 3nRR can produce mature gametes, whereas the male 3nRR cannot. In addition, we produced diploid (2nF 1 , 2n = 100), triploid (3nF 1 , 3n = 150) and tetraploid (4nF 1 , 4n = 200) hybrids in the F 1 generation by crossing females of 3nRR with males of RCC, which further indicated that female 3nRR were fertile. Gonadal transcriptome reveals 6 hub genes (MYC, SOX2, BMP4, GATA4, PTEN and BMP2) were involved in the female fertility of the female 3nRR, and 2 hub genes (TP53 and FGF2) were involved in the male sterility of the male 3nRR. The obtained data reveals novel candidate genes for the fertility in the autotriploid fish and also extends an understanding of the molecular aspects of fertility in triploid fish.

Animals and crosses
One year old RCC and one year old 4nRR (F 11

Gonadal histologic analysis
Ploidy levels of the fish (RCC, 4nRR and 3nRR) were estimated using a flow cytometer (Gallios Flow Cytometer, Beckman Coulter). Blood was collected from the caudal vein using heparinized syringes. Samples were then resuspended in 4,6-Diamidino-2-Phenylindole solution (Sigma-Aldrich) for 10 min. DNA content was compared with that of RCC per sample. Gonadal tissues of two years old female RCC, male 4nRR and 3nRR were fixed in Bonn's liquid and then dehydrated using graded series of alcohol, cleared with xylene, embedded in paraffin wax and cut into 5-8 μm sections. The sections were placed on slides, stained with hematoxylin and eosin, and viewed under a light microscope.

Gamete phenotypes and egg ploidy detection
The water-like semen and mature eggs of two years old 3nRR were sampled for morphological examination. The female 3nRR produced different sized eggs. To determine the egg ploidy, mature eggs were used in vitro fertilization of the RCC haploid sperm and then viable offspring (2nF 1 , 3nF 1 and 4nF 1 ) were generated. Ploidy of these offspring was detected by chromosome counts.

Preparation of chromosome spreads
For ploidy level analysis, chromosome counts were carried out using kidney tissues from 10 individuals each of RCC, 4nRR, 3nRR, 2nF 1 , 3nF 1 and 4nF 1 at eight months of age following a previously described method [67]. 200 metaphase chromosome spreads (20 spreads per sample) were analyzed for each type of fish. Each preparation was examined under 3330× magnification with an oil immersion lens.

Declarations
Ethics approval and consent to participate The study was approved by Ethics Committee of Hunan Normal University, all methods were carried out in accordance with relevant guidelines and regulations. This study was carried out in compliance with the ARRIVE guidelines.

Consent for publication
Not applicable.