Dynamic Transcriptome Profiling Reveals LncRNA-Centred Regulatory Networks in the Modulation of Pluripotency

Abstract

Long noncoding RNAs (lncRNAs) have emerged as vital regulators of gene expression during embryonic stem cell (ESC) self-renewal and differentiation. Here, we systemically analyzed the differentially regulated lncRNAs during ESC-derived cardiomyocyte (CM) differentiation. We established a perspicuous profile of lncRNA expression at four critical developmental stages and found that the differentially expressed lncRNAs were grouped into six distinct clusters. The cluster with specific expression in ESC enriches the largest number of lncRNAs. Investigation of lncRNA-protein interaction network revealed that they are not only controlled by classic key transcription factors, but also modulated by epigenetic and epitranscriptomic factors including N⁶-methyladenosine (m⁶A) effector machineries. A detailed inspection revealed that 28 out of 385 lncRNAs were modified by methylation as well as directly recruited by the nuclear m⁶A reader protein Ythdc1. Unlike other 27 non-coding transcripts, the ESC-specific lncRNA Gm2379, located in both nucleus and cytoplasm, becomes dramatically upregulated in response to the depletion of m⁶A or Ythdc1. Consistent with the role of m⁶A in cell fate regulation, depletion of Gm2379 results in dysregulated expressions of pluripotent genes and crucial genes required for the formation of three germ layers. Collectively, our study provides a foundation for understanding the dynamic regulation of lncRNA transcriptomes during ESC differentiation and identifies the interplay between epitranscriptomic modification and key lncRNAs in the regulation of cell fate decision.

Keywords: Gm2379; cell fate; embryonic stem cell; lncRNA; m6A modificaiton.

FIGURE 1

Dynamic lncRNA expression pattern during ESC-derived cardiac differentiation. (A) Integrative Genomics Viewer (IGV) of read coverage for specific marker genes at distinct differentiation phases including ESC (Day 0), MES (Day 4), CP (Day 6) and CM (Day 10). The number of reads indicate the peak heights in corresponding genomic regions. (B) Genomic annotation of lncRNAs (n = 6,061) showing that the majority of lncRNAs were mapped to lincRNA (50.06%) and antisense (34%). (C) Heat map showing the expression patterns of lncRNAs were grouped into six different groups (C1-C6) across four differentiation stages. (D) Clustering analysis showing the expression patterns of lncRNAs were divided into six different groups (C1-C6) as shown in (C). The average trend is indicated by black lines. The number of genes is indicated for each cluster. The ESC-specific cluster (C5) contains the largest population of lncRNAs (n = 385).

FIGURE 2

Regulatory features and cellular functions of ESC-specific lncRNAs. (A) Length distribution of ESC-specific lncRNA transcripts. (B) Polyadenylation features of ESC-specific lncRNAs. (C) Motif enrichment analysis showing representative TF motifs identified in the promoter region of genes coding ESC-specific lncRNAs. The selected motifs are linked to their corresponding TF. (D,E) Bubble plots show the biological pathways (D) and molecular functions (E) enriched by ESC-specific mRNAs co-expressed with lncRNAs. The size of a bubble is proportional to the number of genes. The intensity of the circle color is corresponding to the significance level which is indicated as -log₁₀ (adjust p-value).

FIGURE 3

Identification of ESC-specific lncRNA-binding proteins. (A) GO analysis of molecular functions enriched by identified binding proteins for ESC-specific lncRNAs. (B) The lncRNA-protein interaction network identifies the candidate proteins binding with ESC-specific lncRNAs Xist, Gas5 or Sox1/2ot. (C) The lncRNA-protein interaction network identifies the candidate lncRNAs binding with pluripotency regulators Oct4, PRC2 and m⁶A machinery proteins. Yellow and green respectively represent lncRNAs and the interacting proteins of lncRNAs. See also Supplementary Table S2.

FIGURE 4

Characterization of the modification status on ESC-specific lncRNAs by m⁶A. (A) Transcriptome-wide distribution pattern of m⁶A for ESC-specific m⁶A-containing lncRNAs. (B) The percentage and number of ESC-specific lncRNAs containing m⁶A peaks that are regulated by Mettl3. (C) A Venn diagram shows the overlapping of lncRNA transcripts containing m⁶A sites and binding with nuclear m⁶A “reader” Ythdc1. (D) A box plot shows the expression of 28 overlapped lncRNAs (C) containing m⁶A sites and binding with Ythdc1 in response to the depletion of Mettl3 or Ythdc1. (E) A scatter plot showing the expression changes of 28 overlapped lncRNAs (C) in the presence or absence of Ythdc1. (F) A scatter plot showing the expression changes of 28 overlapped lncRNAs (C) in the presence or absence of Mettl3. Gm2379 is highlighted in green.

FIGURE 5

The m⁶A-harboring lncRNA Gm2379 is regulated by both Mettl3 and Ythdc1. (A) Detection of Gm2379 transcript levels during ESC-derived cardiac differentiation by RT-qPCR. Nanog and Nkx2-5 serve as pluripotent and cardiac maker genes. Error bars, mean ± SEM; n = 4 independent replicates; *p < 0.05, **p < 0.01, ***p < 0.001. (B) Detection of the distribution of Gm2379 transcript in ESCs by cell fractionation assay. The RT-qPCR data represented a percentage of the total amount of detected transcripts. Error bars, mean + SEM; n = 2 independent replicates. (C) Genome browser view of a representative genomic region showing m⁶A read density and Ythdc1-binding feature of the Gm2379 transcript. (D) Detection of Gm2379 transcript levels in WT and Mettl3-KO ESCs by RT-qPCR. Error bars, mean + SEM; n = 3 independent replicates; **p < 0.01. (E) Detection of Gm2379 transcript levels in control and Ythdc1-KO ESCs by RT-qPCR. Error bars, mean + SEM; n = 4 independent replicates; **p < 0.01.

FIGURE 6

Gm2379 modulates the expressions of pluripotent and germ-layer genes. (A) Generation of Gm2379-KD ESCs. (B) Detection of transcript levels of pluripotent genes in control and Gm2379-KD ESCs in the presence or absence of LIF by RT-qPCR. (C) Detection of transcript levels of three germ-layer genes in control, Gm2379-KD and Mettl3-KO ESCs by RT-qPCR. shGm2379-1 ESCs were used in (B) and (C). Error bars, mean + SEM; n = 4 independent replicates; *p < 0.05, **p < 0.01, ***p < 0.001.