Chromatin Remodeling via Retinoic Acid Action during Murine Spermatogonial Development

Abstract

Spermatogonial differentiation is a process that commits germ cells to the complex process of spermatogenesis. Spermatogonial differentiation is mediated by the action of retinoic acid, which triggers major morphological and transcriptional changes. While these transcriptional changes have been well explored, there has been little effort devoted to epigenetic regulation surrounding spermatogonial development. This study aimed to uncover the timing and dynamics of chromatin organization during spermatogonial development within the context of these transcriptional changes. Using germ cell synchrony and the assay for transposase accessible chromatin and next generation sequencing (ATAC-seq) to isolate subpopulations of developing spermatogonia and identify accessible regions within their genome, we found that 50% of accessible regions in undifferentiated spermatogonia were condensed following retinoic acid action within 18 h. Surprisingly, genes with known functional relevance during spermatogonial development were accessible at all times, indicating that chromatin state does not impact transcription at these sites. While there was an overall decrease in gene accessibility during spermatogonial development, we found that transcriptionally active regions were not predictive of chromatin state.

Keywords: ATAC-seq; chromatin; retinoic acid; spermatogenesis; testis.

Figure 3

Peak variance relative to transcript regulation. Number of peaks variance shown relative to the mean for genes with significantly (A) decreased or (B) increased transcript counts by 120 h post-RA relative to 0 h post-RA. List of genes’ number of peaks per timepoint and the plotted z-Scheme 5 (Figure 4 and Figure S5). The two most common categories for chromatin accessibility were genes accessible at all timepoints, or those only accessible at 0 h post-RA (Figure 2). For those genes only accessible at 0 h, there was very little variation in transcription across spermatogonial development (Figure 4A). The same trend was seen for those genes which were accessible at all timepoints (Figure 4B). Thus, those genes which were accessible only prior to RA administration and those accessible at all timepoints showed similar transcript expression patterns.

Figure 1

Accessible regions. (A) The shared accessible regions between the two replicates at each timepoint for 0, 18, 48, and 120 h post-RA revealed a monotonic decrease in accessibility following RA treatment. (B) Unique accessible regions for each comparison of times post-RA treatment. (C) Accessible regions grouped by genomic traits for each replicate and timepoint. Notably, the total regions here may show a greater total than Figure 1A, as these represent regions called in a single replicate not only those which were accessible in both replicates of a timepoint.

Figure 2

Accessible genes. Genes with associated accessible regions at 0, 18, 48, and/or 120 h post-RA showed that 72.8% of accessible genes were accessible at all timepoints. The numbers in the table show the number of regions per category, and percentages of the total are shown to the right of the legend. Genes with accessible regions were included for a timepoint if both biological replicates indicated that the gene contained and/or was within 10 kb of an accessible region.

Figure 4

Accessible gene and transcript values. Heatmaps of transcript values during spermatogonial development for genes which have accessible chromatin at (A) 0 h only or (B) 0, 18, 48, and 120 h (all times). Transcript counts and plotted z-scores are provided in the Supplementary Materials. Transcript data from [5].

Figure 5

Accessibility of classical markers. Relative sequencing tracks the mapping accessibility of common spermatogonial genes and immunohistochemical markers. Numbered carats denote significant accessibility peaks for each gene. Data shown for both replicates of each timepoint: 0 h (yellow and orange), 48 h (light and dark green), and 120 h (light and dark blue). The genes represented here include (A) Nanos2, (B) Lin28a, (C) Neurog3, (D) Pou5f1, (E) Sox3, and (F) Zbtb16, all of which show a decreasing amplitude and width of peaks between the undifferentiated and differentiating spermatogonia corresponding with decreasing accessibility.

Figure 6

Accessibility of uncharacterized undifferentiated spermatogonia genes. Accessibility of uncharacterized genes, which are highly expressed at the 0 h timepoint but then decrease 18 h post-RA [5]. Numbered carats denote significant accessibility peaks for each gene. Relative sequencing tracks the mapping accessibility of these uncharacterized genes including (A) Enc1 (B) Onecut2 (C) Ptch1 (D) Sdc4, and (E) Zfp462. The timepoints and each of their replicates are represented in different rows and by color: 0 h (grey and black), 18 h (yellow and orange), 48 h (light and dark green), and 120 h (light and dark blue).

Figure 7

Accessibility of functional spermatogonia genes. Relative sequencing tracks mapping accessibility of common spermatogonial genes (A) STRA8 and (B) KIT, both of which show changes in peak amplitude and width between the undifferentiated and differentiating spermatogonia. Numbered carats denote significant accessibility peaks for each gene. The timepoints and each of their replicates are represented in different rows and by color: 0 h (grey and black), 18 h (yellow and orange), 48 h (light and dark green), and 120 h (light and dark blue).