Genome-wide analysis and functional prediction of long non-coding RNAs in mouse uterus during the implantation window

Abstract

Establishment of the receptive uterus is a crucial step for embryo implantation. In this study, the expression profiles and characterization of long non-coding RNAs (lncRNAs) in pregnant mouse uteri on day 4, day 5 at implantation sites and inter-implantation sites were conducted using RNA-seq. A total of 7,764 putative lncRNA transcripts were identified, including 6,179 known lncRNA transcripts and 1,585 novel lncRNA transcripts. Bioinformatics analysis of the cis and trans lncRNA targets showed that the differentially expressed lncRNAs were mainly involved in tissue remodelling, immune response and metabolism-related processes, indicating that lncRNAs could be involved in the regulation of embryo implantation. We also discovered that differentially expressed lncRNAs might regulate multiple signalling pathways that play an important role in the regulation of embryo implantation. In addition, nine known lncRNAs and four novel lncRNAs were randomly selected and validated by qRT-PCR. The expression of Tug1, Neat1, Gas5, Malat1, H19 and Rmst were significantly regulated in the mouse uterus during the implantation window. Our results are the first to systematically identify lncRNAs in the mouse uterus and provide a catalogue of lncRNAs for further understanding their functions in pregnant mouse uteri during the implantation window.

Keywords: RNA-seq; implantation; long non-coding RNAs; mouse; uterus.

Figure 1. LncRNA characteristics in mouse uterus during the implantation window

(A) Expression level (FPKM) comparison between lncRNA and protein-coding genes. The FPKM distribution of lncRNAs in mouse uterus is lower than that of protein-coding genes. (B) Conservation analysis was evaluated using phyloP (http://compgen.bscb.cornell.edu/phast/). The level of conservation of lncRNAs in mouse uterus is lower than that of protein-coding genes. (C) Distribution of transcript lengths in the lncRNAs and protein-coding genes. Transcript size distributions of lncRNAs is generally shorter than that of the protein-coding genes. (D) The exon number in lncRNAs and protein-coding genes. 88.41% of the lncRNAs contains two to four exons, while the majority of protein-coding genes consist of more than 10 exons. (E) The number of ORFs identified in the lncRNAs and protein-coding genes using Estscan. As expected, the ORFs of lncRNAs is substantially shorter than that of protein coding genes. (F) Subtypes of the putative lncRNAs according to the latest gene/transcript biotypes in GENCODE & Ensembl.

Figure 2. Venn diagrams of the differentially expressed lncRNA transcripts in three comparison groups during the implantation window

Differential expression analysis was performed using the Cuffdiff program basing on FPKM value derived from Cufflinks, and p-values less than 0.05 were considered as significantly differentially expressed between two groups.

Figure 3. Hierarchical clustering analysis of differentially expressed lncRNAs in mouse uterus during the implantation window

(A) The heat map of differentially expressed lncRNAs in the three comparison groups. Red indicates higher expression and blue indicates lower expression. Enrichment analysis of GO terms for differentially expressed lncRNAs in cis (B) and trans (C) from the three comparison groups. Red and green show higher and lower expression, respectively.

Figure 4. qRT-PCR validation of 13 differentially expressed lncRNAs in RNA-seq data

The mouse uterine sample on pregnant day 1 (D1U), day 4 (D4U), day 5 at implantation sites (D5IU) and inter-implantation sites (D5NU) and day 8 at implantation sites (D8IU) were collected, respectively. LncRNAs expression was normalized with GAPDH using the 2^−ΔΔCt analysis method. The data are expressed as the mean ± S.E.M. from three replicates, and the bar bearing different superscript indicates a significant difference between the mean values (P < 0.05).