Systematic identification and characterization of long non-coding RNAs in mouse mature sperm

Abstract

Increasing studies have shown that mature spermatozoa contain many transcripts including mRNAs and miRNAs. However, the expression profile of long non-coding RNAs (lncRNAs) in mammalian sperm has not been systematically investigated. Here, we used highly purified RNA to investigate lncRNA expression profiles in mouse mature sperm by stranded-specific RNA-seq. We identified 20,907 known and 4,088 novel lncRNAs transcripts, and the existence of intact lncRNAs was confirmed by RT-PCR and fluorescence in situ hybridization on two representative lncRNAs. Compared to round spermatids, 1,794 upregulated and 165 downregulated lncRNAs and 4,435 upregulated and 3,920 downregulated mRNAs were identified in sperm. Based on the "Cis and Trans" RNA-RNA interaction principle, we found 14,259 targeted coding genes of differently expressed lncRNAs. In terms of Gene ontology (GO) analysis, differentially expressed lncRNAs targeted genes mainly related to nucleic acid metabolic, protein modification, chromatin and histone modification, heterocycle compound metabolic, sperm function, spermatogenesis and other processes. In contrast, differentially expressed transcripts of mRNAs were highly enriched for protein metabolic process and RNA metabolic, spermatogenesis, sperm motility, cell cycle, chromatin organization, heterocycle and aromatic compound metabolic processes. Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis showed that the differentially expressed lncRNAs were involved in RNA transport, mRNA surveillance pathway, PI3K-Akt signaling pathway, AMPK signaling pathway, protein processing in endoplasmic reticulum. Metabolic pathways, mRNA surveillance pathway, AMPK signaling pathway, cell cycle, RNA transport splicesome and endocytosis incorporated with the differentially expressed mRNA. Furthermore, many lncRNAs were specifically expressed in testis/sperm, and 880 lncRNAs were conserved between human and mouse. In summary, this study provides a preliminary database valuable for identifying lncRNAs critical in the late stage of spermatogenesis or important for sperm function regulation, fertilization and early embryo development.

Fig 1. Quality control of sperm RNA.

Confirmation of sperm purity and electrophoretic size distribution of extracted total sperm RNA. (A) Morphology of collected sperm under phase contrast microscopy. (B) Biomarkers of leukocytes (CD4), testicular germ cells (c-kit) and epithelial cells (E-cadherin) could be amplified easily from the RNA extracted from unpurified sperm sample. (C) After sperm was purified, non-sperm cell markers (CD4, c-kit, E-cadherin) were unable to be amplified from the RNA extracted from our purified sperm sample, while positive markers of sperm (Prm1 and Prm2) were easily detected. (D, E) The 442 bp PCR product of Prm2 indicated DNA contamination (D), which disappeared after DNase I digestion (E). (F) The DNA PCR product of Prm1 and Prm2 could be amplified from no reverse transcription (RT) sperm RNA. (G, H) Electrophoretic size distribution of RNAs in mouse testis (G) and mature sperm (H) analyzed by Agilent Bioanalyzer.

Fig 2. LncRNAs expression profiles in mouse mature sperm.

(A) Correlation analysis of all sequencing samples. (B) The numbers of lncRNAs according to length in mice testes and mature sperm. (C) The numbers of lncRNAs and mRNAs in mature sperm. (D) The average expression levels of known and novel lncRNAs and mRNAs in mice mature sperm. (E) The average expression levels of known and novel lncRNAs in mice testes and mature sperm. (F) The violin graph of lncRNAs and mRNAs in in mice testes and mature sperm. **P < 0.01 compared with mRNA, ##P < 0.01 compared with testes.

Fig 3. Integrity evaluation and existing forms of sperm lncRNAs.

(A) Integrity evaluation of sperm lncRNAs using a computational approach. To determine the uniformity of coverage for each, RNA transcripts were divided into 100 bins, and a 5-bin moving average was used to calculate localized variations in sequencing coverage. The squared deviation from expected coverage for each bin was summed and used as an intactness score to rank the 1000 lncRNAs according to their stability. (B) Intact RNAs are present in sperm; the testes served as positive control to amplify the proposed full-length mRNAs and lncRNAs. (C) The fragmented RNAs degraded or cleaved from the intact transcripts were amplified in sperm. S′ and T′: Representative fragments were amplified from sperm and testes cDNA, respectively. (D) The localization of two representatively intact lncRNAs with FISH staining.

Fig 4. Validation of tissue-specific expression of lncRNA by RT-PCR.

Total RNA of testis, lung, liver, stomach, brain, large intestine, spleen, fat, muscle, kidney, and sperm were isolated using RNeasy® Plus Micro kit, and RT-PCR was performed to detect the lncRNA candidates expression level. Rplp1 genes were used as loading controls.

Fig 5. The putative functional lncRNAs in mature sperm.

LncRNA target genes were predicted based on the “Cis and Trans” RNA-RNA interaction principle. One-to-one pairs of lncRNA and mRNA were deemed to putative functional lncRNAs modulating the spermatogenesis or sperm function.

Fig 6. qPCR validation of the upregulated RNAs.

After total sperm RNA was isolated, qPCR was performed to detect the RNA expression in sperm and testes. Rplp1 and β-actin genes were used as loading controls to normalize RNA expression levels. Data are expressed as the mean ± standard deviation (n = 3).