Fish broodstock and gamete collection

Gametes were collected from 3- and 4-year-old individuals of spring spawning rainbow trout (Oncorhynchus mykiss Walbaum 1792) coming from the outbred selection strain (Rutki strain) cultured at the Department of Salmonid Research, Inland Fisheries Institute in Olsztyn, Rutki, Poland. Rainbow trout from the Rutki strain exhibit chromosomal polymorphism of the Robertsonian type and their diploid (2n) chromosome number ranges from 59 to 62, whereas the chromosome arm number value (FN) is stable and equals 10441.

Before manipulations, selected specimens were anesthetized using Propiscin (etomidate, IRŚ, Poland) at a dose of 0.5 ml l−1 of water. Eggs from four females were stripped and collected into separate polyethylene (PE) bowls. Eggs from each female (c. 3700) were then divided into three batches to produce haploid androgenetic (Andro 1n) (A), gynogenetic (Gyno 1n) (G) and diploid control (C) groups and transferred to separate plastic containers with covers, submerged in the ovarian fluid and kept at 8–10 °C for further treatment. Milt from four males was collected into the separate PE containers. To confirm the sperm motility, 1 μl of semen with an addition of 49 μl of the sperm activating medium (SAM; 154 mM NaCl and 1 mM Ca2+, buffered to pH 9.0 with 20 mM Tris + 30 mM glycine)42 was placed on a glass slide and analyzed under a microscope (Nikon Eclipse E 2000). Sperm from all striped males showed high (>80%) motility. Sperm samples were refrigerated at 2 °C for further use.

Inactivation of the parental nuclear DNA

Batches of eggs to be used for androgenesis were placed in a cooler box and transported to the Clinic of Oncology and Radiotherapy, University Clinical Center, Medical University of Gdansk. A standard irradiation method was used to inactivate the nuclear DNA in rainbow trout eggs9,43. A TrueBeam linear accelerator (Varian Medical Systems, Palo Alto, CA, USA) was used to irradiate eggs [350 Gy of X-rays (6 Gy/min)]. Eggs were irradiated from two opposed fields (from top and from bottom) with a dose of 175 Gy from each field. The distance from the radiation source was 98.7 cm. During irradiation that lasted a. 66 min, containers with eggs immersed in ovarian fluid that formed a layer of 26 mm were placed on a PE tray with crushed ice.

Sperm from four rainbow trout males was pooled and diluted in the rainbow trout seminal plasma (40×). A 60 ml glass beaker (50 mm diameter, 30 mm height) with 15.375 ml of diluted sperm (depth of diluted sperm: 7.8 mm) was placed on a magnetic stirrer and exposed to the UV-C light source (Phillips TUW 30 Watt UV bulb) for 11 min. The distance between the surface of the magnetic stirrer and the UV lamp was 20 cm. The UV intensity and UV dose were 2075 μW/cm2 and 1369.5 J/m2, respectively. During the irradiation, the diluted sperm was mixed using the magnetic stirrer (1400 × g). The irradiated sperm was used for the gynogenetic egg activation immediately after the UV exposure.

Induction of haploid androgenesis and gynogenesis

One hour after irradiation, batches of irradiated eggs were separately inseminated with spermatozoa from each male in the presence of SAM43 to provide androgenetic haploids (Andro 1n)11,43. Portions of diluted and UV-irradiated sperm were added separately to batches of non-irradiated eggs from four females (~150,000 spermatozoids per egg) and poured over with SAM to induce gynogenetic development (Gyno 1n). To provide diploid control groups (C), non-irradiated eggs were inseminated separately with milt from each male.

Fertilized eggs were placed in a hatching apparatus and incubated at 9 °C at the Department of Salmonid Research, Rutki. The survival of normal diploid and haploid androgenetic and gynogenetic embryos were calculated 27 days post fertilization (dpf) (eyed stage) and 57 dpf (swim-up stage).

For the transcriptome analysis, portions of about 50 non-activated eggs from each batch/female were taken and transferred into 50 ml plastic falcons with RNA-later solution (Sigma Aldrich) before irradiation (non-irradiated eggs) (1), after irradiation with 175 Gy (2), and after irradiation with 350 Gy of X rays (3). Moreover, fertilized (activated) irradiated eggs (Andro 1n) (n = 50) (4) and activated non-irradiated eggs (Gyno 1n) (n = 50) (5) from each female were taken and submerged in the RNA-later solution 4 h after insemination (5 h after egg irradiation) (4).

Cytogenetic examination of embryos

Androgenetic and normal diploid rainbow trout eyed embryos were cytogenetically studied using the in vivo technique7 to confirm their haploid state. For the purpose of visualization, chromosomes were stained with 4′,6-diamidino-2-phenylindole DAPI (Vector, Burlingame, USA) and examined under a Zeiss Axio Imager A1 microscope equipped with a fluorescent lamp and a digital camera (Applied Spectral Imaging, Galilee, Israel). Images were captured and processed using the Band View/FISH View software (Applied Spectral Imaging).

RNA quantity and degradation analysis

Eggs preserved in RNA-later solution were incubated overnight in a refrigerator (4 °C) and finally stored at −80 °C for RNA purification. To assess the relative RNA yield in eggs from each female, four eggs of similar size from each batch [non-irradiated eggs, eggs irradiated with 175 Gy, eggs irradiated with 350 Gy, fertilzied irradiated eggs (Andro 1n), and gynogenetically activated non-irradiated eggs (Gyno 1n)] were thawed on ice and homogenized in TRIzol Reagent (Thermo Fisher Scientific) using a manual method to avoid homogenization-induced mRNA degradation. RNA was purified using the modified TRIzol procedure established at the Igor Babiak laboratory, University of Nordland, Norway. The obtained RNA was suspended in the equal volumes and quantified using three different methods: (i) the spectrophotometric method – with a NanoDrop2000 spectrophotometer (Thermo Fisher Scientific), (ii) the fluorometric method – using RNA binding dye and a Qubit 2.0 Fluorometer (Thermo Fisher Scientific) and (iii) the electrophoretic method with RNA-binding dye using the Agilent 2200 TapeStation system (RNA screen tapes). For each quantification, the RNA purification was carried out in separate rounds of preparation to account for bias associated with egg size and random sampling. Separate rounds of RNA purification also allowed to account for the effect of some potential small differences in the course of the manual RNA isolation protocol. An Agilent’s RIN (RNA Integrity Number) algorithm was also used to compare the RNA quality/integrity between separate treatments. The algorithm analyzes the ratio of 18S and 28S rRNA and the whole electrophoretic trace of a sample, which indicates the presence or absence of degradation signs. The assigned RIN is independent of the sample concentration and assigns a value of 1 to 10 to an electropherogram, with 10 being the highest quality RNA. RIN is standardized and can be used for comparative studies44.

To evaluate the impact of irradiation on the quantity and quality of the maternal mRNA from rainbow trout eggs, the analyzed parameters were compared between different groups using the paired and unpaired t-test.

Whole transcriptome sequencing and data analysis

To provide the highest RNA purity needed for sequencing, the total RNA was additionally purified before the library construction, applying Agencourt RNAClean XP beads (Beckman Coulter). 800 ng of purified RNA was used for the TruSeq RNA Sample Prep v2 kit (Illumina). Library construction process included mRNA selection, fragmentation, cDNA synthesis, end repair, adenylation, indexed adapter ligation and amplification and it was followed by quantitative (Qubit, Thermo Fisher Scientific) and qualitative (Agilent TapeStation 2200) assessment. Quality controlled and normalized libraries were ultimately sequenced in a single 50-bp run (1 × 50 bp) using the HiScanSQ system with application of the TruSeq SBSv3 Sequencing kit (Illumina) to achieve approx. 25 million reads per sample.

Achieved demultiplexed raw reads were controlled for quality using the FastQC software and filtered using the Flexbar software44 in order to trim a random sequence of adapters and remove low quality reads. The obtained high quality reads were mapped against the newest accessible rainbow trout transcriptome atlas (encompassing 44 990 transcripts; downloaded from, using Bowtie software45, the set for unlimited multi-mappings (−a). The mapped reads were studied using the eXpress software ( Roberts and Pachter, 2012), which was designed to estimate the number of transcripts in multi-isoform genes and was able to resolve multi-mappings of reads across gene families. As the software does not need a reference genome, it may be used with de novo transcriptome assemblies. The implemented model is based on the formerly described probabilistic models established for RNA-Seq.47. Nevertheless, it is also appropriate to other settings where target sequences are sampled and contain values for fragment size distributions, errors in reads and a fragment sequence bias48. The achieved assessed rounded effective read counts for separate transcripts and samples were calculated using DESeq249. The data analysis steps followed guidelines presented by Roberts and Pachter46 in the eXpress software guide.

The functional annotation and gene ontology (GO) analysis of differentially expressed transcripts was performed with the Blast2GO software50.

miRNA library preparation, sequencing and data analysis

For miRNA analysis, RNA was purified using an in-house modified procedure with the use of the Direct-zol RNA Mini Prep kit (Zymo Research). In brief, eggs were manually homogenized in TRIzol solution and subjected to phase-separation with chloroform according to a protocol established at Igor Babiak Laboratory (University of Nordland, Bodo, Norway)34. The upper phase and 99.8% EtOH were mixed (1:1) and from this point, the Direct-zol RNA Mini Prep protocol was followed. The extracted RNA was assessed in terms of its quantity and quality, with the use of a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific) and a TapeStation 2200 instrument (Agilent), respectively. The amount of 500 ng of RNA was used to construct microRNA libraries using the NEBNext Multiplex Small RNA Library Prep Set for Illumina (New England Biolabs) and following the standard protocol with the multiplexing option (3′ adaptor ligation, Reverse Transcription Primer ligation, 5′ adaptor ligation, PCR amplification, PAGE gel purification and EtOH precipitation). The libraries were subjected to concentration measurements using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific) and size measurements with a 2200 TapeStation instrument (Agilent). The normalized miRNA libraries were finally sequenced with the PhiX control library (Illumina) in a single 50-bp run on the HiScanSQ system using the TruSeq SBSv3 Sequencing kit (Illumina).

The acquired reads were controlled for quality using the FastQC software51 and then analyzed with the use of UEA sRNA Workbench V3.252. First, 3′ adaptor sequences were trimmed off and the sequences were filtered according to the following parameters: 17–35 nt in length, minimum abundance of at least 6 supporting reads, tRNA and rRNA sequences discarded from the data. The microRNA sequences were identified with reference to the rainbow trout genome ( and miRBase v21.0 Atlantic salmon Salmo salar homologues53,54, applying the miRCat tool with default parameters for animals with the exception of minimum length (17 nt), maximum length (25 nt) and minimum abundance (6). In addition, potentially novel miRNA sequences were checked for the presence of other non-coding RNA species, using the RNAcentral database (the RNAcentral Consortium, 2017). In the last step, the microRNA length and sequence variants (isomiRs) were identified using the isomiR-SEA software55 with the default settings.

The differential expression analysis of the identified miRNAs was performed using the DESeq2 software56 following the software manual. The miRNAs showing statistically significant differences between the tested samples (p ≤ 0.05) were selected for further analysis.

The differentially expressed miRNAs were analyzed using the mirPath v.3 DIANA Tools web application57 to identify their target genes and enriched pathways. We chose Tarbase v 8.0 and microT-CDS bases as reference databases of miRNA target genes, while the GO database was employed for the pathway enrichment analysis. As rainbow trout microRNAs are not present in the miRBase (21.0), human and zebrafish Danio rerio homologues deposited in miRBase (21.0) were used.

qPCR validation of RNA-Seq results

For qPCR validation, two genes: the immediate early response 2 (IER2) gene and the early growth response 1 (EGR1) gene, differentially expressed in the irradiated eggs, were selected. The primers for the amplification of the gene fragments (Supplementary Table 4) were designed with the use of the Primer3 software within the transcript regions present in all splicing variants found in the employed version of the transcriptome. The EF1-α (Translation elongation factor 1-alpha) gene was selected as an endogenous control and amplified using primers previously described by Bland et al.58. This gene had a stable transcript level in all analyzed groups in the present study as well as it was previously shown to be an useful endogenous control in several fish species58. The purified total RNA was transcribed into cDNA using the High-Capacity RNA-to-cDNA™ Kit (Thermo Fisher Scientific, MA, USA). Expression levels were determined with the QuantStudio 7 Flex System (Thermo Fisher Scientific, Waltham, MA, USA) using PowerUp SYBR™ Green Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) and the standard manufacturer protocol. The specificity of primers was confirmed by melting curve analysis and agarose gel electrophoresis of PCR products. The qPCR reactions with cDNA samples from non-irradiated eggs, eggs irradiated with 175 Gy and 350 Gy were conducted in triplicates with corresponding negative controls. The efficiency of the amplification reaction for different pairs of primers was determined using a standard curve method. The relative expression ratio was calculated according to Pfaffl59. Expression fold changes obtained from RNA-Seq and qPCR were compared. In addition, correlation coefficients between the results of RNA-Seq and qPCR were calculated for the expression levels established in individual samples as well as for the group means (Supplementary File 12).

Approval (Ethical Committee for the Experiments on Animals)

This study was performed in compliance with the recommendations contained in the Polish Act of 21 January 2005 on Experiments on Animals (Dz. U. of 2005, No. 33, item 289). The protocol was approved by the Local Ethical Committee for the Experiments on Animals of the University of Gdansk, Poland (no. 28/2015). This article does not cover any studies with human participants, performed by any of the authors.


All methods applied in the present paper were carried out in accordance with the relevant guidelines and regulations.