Expression dynamics, relationships, and transcriptional regulations of diverse transcripts in mouse spermatogenic cells

Abstract

Among all tissues of the metazoa, the transcritpome of testis displays the highest diversity and specificity. However, its composition and dynamics during spermatogenesis have not been fully understood. Here, we have identified 20,639 message RNAs (mRNAs), 7,168 long non-coding RNAs (lncRNAs) and 15,101 circular RNAs (circRNAs) in mouse spermatogenic cells, and found many of them were specifically expressed in testes. lncRNAs are significantly more testis-specific than mRNAs. At all stages, mRNAs are generally more abundant than lncRNAs, and linear transcripts are more abundant than circRNAs. We showed that the productions of circRNAs and piRNAs were highly regulated instead of random processes. Based on the results of a small-scale functional screening experiment using cultured mouse spermatogonial stem cells, many evolutionarily conserved lncRNAs are likely to play roles in spermatogenesis. Typical classes of transcription factor binding sites are enriched in the promoters of testis-specific m/lncRNA genes. Target genes of CREM and RFX2, 2 key TFs for spermatogenesis, were further validated by using ChIP-chip assays and RNA-seq on RFX2-knockout spermatogenic cells. Our results contribute to the current understanding of the transcriptomic complexity of spermatogenic cells and provide a valuable resource from which many candidate genes may be selected for further functional studies.

Keywords: circRNAs; expression; lncRNAs; mRNAs; spermatogenesis; testis; transcription.

Figure 1.

Expression of mRNAs, lncRNAs and circRNAs in primitive type A spermatogonia (priSG-A), preleptotene spermatocytes (plpSC), pachytene spermatocytes (pacSC), round spermatids (rST), collectively referred to as 4 spermatogenic cell (SPC) types. (A) Frequency distributions of the log₂-transformed FPKM values of mRNAs and lncRNAs expressed in the 4 SPC types. (B) Distributions of the relative abundances of circRNAs to linear RNAs evaluated in log₂-transformed ratios of junction reads representing circRNAs over the junction reads representing linear RNAs in the 4 SPC types and cultured spermatogonial stem cells (SSC) as well as in cerebellum and cortex, the RNA-seq data of which were reported by Rybak-Wolf et al. (C, D) Distributions of the relative abundances of circRNAs overlapping (C) or not overlapping (D) those detected in the RNase R treated SSC sample (SSC-R⁺). (E) Identity validation of circRNAs by RT-PCR using 2 reverse transcriptases in 2 independent experiments (upper panel: MMLV-derived reverse transcriptase; lower panel: AMV-derived reverse transcriptase). The circRNAs were represented by the names of the corresponding genes. The circled numbers on the right side mark the subsets, from which the circRNAs were selected: 1, sets shown in (C); 2 and 3, sets corresponding to the left and right peaks of the plots in (D), respectively. (F) Clusterings of m/lnc/circRNAs expressed in the 4 SPC types, which are named g-m/lnc/circRNAs, respectively. (G) Expression validation of lncRNAs from the 3 clusters (sg-lncRNAs, sc-lncRNAs and st-lncRNAs) by qRT-PCRs. The expression levels were represented by delta-CT values. The expression of 4 well-known protein coding mRNAs, which are differentially expressed in the 4 SPC types, were included to show that these mRNAs were expressed with the expected dynamics in the isolated cells (Marker mRNAs). Results for 4 lncRNAs from each cluster were shown. Results for all lncRNAs are shown by Fig. S1.

Figure 2.

Identification of testis-specific m/lnc/circRNAs. (A) RT-PCR valuation of predicted ts-mRNAs. M, marker; Te, testis; Ov, ovary; Ut, uterus; Li, liver; Lu, lung; Mu, muscle; He, heart; Br, brain; Sp, spleen; Ki, kidney. Rik-1*:4932414N04Rik; Rik-2*:1700011F14Rik; Rik-3*:1700019N12Rik; Rik-4*: 1700001C02Rik; Rik-5*: 4930480E11Rik; Rik-6*: 4930407I10Rik. (B) Clustering of all tissue-specific m/lncRNAs (max JS score >0.5). Plots of JS_testis scores were shown to the right of the heat maps. (C) Expression patterns of g-ts-m/lncRNAs in the 4 SPC types. (D) Comparisons of circRNAs and their cognate mRNAs in SPCs and the brain.

Figure 3.

Comparisons of the expression of related RNAs. (A) Expression of different clusters of circRNAs and their cognate linear mRNAs. (B) Expression of piRNAs and their precursor m/lncRNAs. The expression of piRNAs were based on previously published RNA-seq data of small RNAs in 3 types of SPCs. (C) Expression of lncRNAs that map to the exons or introns or the nearby region of mRNA genes and the expression of these mRNAs.

Figure 4.

A small scale screening for functional sg-lncRNAs. (A) qRT-PCR validation of the knockdown (KD) of 6 sg-lncRNAs by siRNAs. (B) The proliferation of the KD SSCs relative to that of the normally cultured SSCs. SSCs were also transfected with siRNAs of scrambled sequence to show that the proliferation was not changed by the transfection reagents. (C) Immunofluorescent images of the KD SSCs to show the appearance of c-KIT-positive cells in some of the KO SSC cultures. The left panel shows the negative staining of c-KIT in controls and 2 lncRNA-KD samples while the right panel shows the positive staining of c-KIT in 4 lncRNA-KD samples. Note the membrane-localized typical staining pettern of c-KIT signal in contrast to the ubiquitous nonspecific background.

Figure 5.

Transcription factor binding sites (TFBSs) enrichment analysis. (A) TF families and members, the TFBSs of which are enriched in the promoters of different sets of ts-mRNAs. Similar TFBSs enriched in each gene set were combined into a family and the lowest p-value of the members was used as the representative p-value of this group. (B) Expression of the CREB and RFX family TFs based on RNA-seq data. (C) ChIP-PCR validation of CREM targets supported by ChIP-chip data (see Table S12 for primer information). (D) Plots of ChIP-chip signals of some genes that were validated by ChIP-PCR in (C). Myog is a muscle specific gene and used as a negative control. The red dots label the signals of the probes on the microarray.