Distinct core promoter codes drive transcription initiation at key developmental transitions in a marine chordate

Abstract

Development is largely driven by transitions between transcriptional programs. The initiation of transcription at appropriate sites in the genome is a key component of this and yet few rules governing selection are known. Here, we used cap analysis of gene expression (CAGE) to generate bp-resolution maps of transcription start sites (TSSs) across the genome of Oikopleura dioica, a member of the closest living relatives to vertebrates. Our TSS maps revealed promoter features in common with vertebrates, as well as striking differences, and uncovered key roles for core promoter elements in the regulation of development. During spermatogenesis there is a genome-wide shift in mode of transcription initiation characterized by a novel core promoter element. This element was associated with > 70% of transcription in the testis, including the male-specific use of cryptic internal promoters within operons. In many cases this led to the exclusion of trans-splice sites, revealing a novel mechanism for regulating which mRNAs receive the spliced leader. During oogenesis the cell cycle regulator, E2F1, has been co-opted in regulating maternal transcription in endocycling nurse nuclei. In addition, maternal promoters lack the TATA-like element found in vertebrates and have broad, rather than sharp, architectures with ordered nucleosomes. Promoters of ribosomal protein genes lack the highly conserved TCT initiator. We also report an association between DNA methylation on transcribed gene bodies and the TATA-box, which indicates that this ancient promoter motif may play a role in selecting DNA for transcription-associated methylation in invertebrate genomes.

Trans-splicing of mRNAs links gene transcription to translational control regulated by mTOR

“…This may be achieved through the use of alternative start sites that either include or exclude the trans-splice site. Supporting this, a male-specific promoter motif recently identified in O. dioica (Danks et al, 2018) causes the exclusion of trans-splice sites, and consequently the lack of a spliced leader on resulting transcripts, in many cases. Given the strong association of trans-splicing with maternal mRNA, mTOR regulation via the spliced leader is likely an underlying mechanism for the adjustment of egg numbers according to nutrient levels in O. dioica (Troedsson et al, 2002).…”

“…The majority of our weakly-translated, Torin 1-resistant, trans-spliced transcripts therefore likely represent dormant maternal mRNAs stocked in oocytes. We used both tiling microarray (Danks et al, 2013) and cap analysis of gene expression (CAGE) (Danks, Navratilova, Lenhard, & Thompson, 2018) data from O. dioica oocytes to determine the set of oocyte-stocked mRNAs. As expected, we found that the translational efficiency of oocyte transcripts in control animals was significantly lower than that of non-oocyte transcripts (mean oocyte log 2 (TE) = -0.80; mean non-oocyte log 2 (TE) = 0.45; Welch two-sample t-test: t = -57.494, df = 14002, P-value < 2.2 x 10 -16 ) ( Figure 3D, Figure S4), and the effect of Torin 1 was significantly reduced (mean oocyte log 2 (Δ) = -0.17; mean non-oocyte log 2 (Δ) = -0.97; Welch two-sample t-test: t = 43.54, df = 12832, P-value < 2.2 x 10 -16 ) ( Figure 3E).…”

“…We wanted to determine whether or not a TOP-like motif is present at the 5' ends of translation-suppressed transcripts that were not trans-spliced. We obtained transcription start sites (TSSs) at bp-resolution in female animals from CAGE data (Danks et al, 2018) and examined the 5' sequences of all expressed transcripts. Out of 2,772 robustly expressed, non-trans-spliced transcripts, only 4 had a canonical TOP motif and only 66 had 5' pyridine-enrichment comparable to the SL sequence.…”

“…We extracted locations of TSSs in day 6 females from an existing CAGE data set (Danks et al, 2018). We used the sequence immediately downstream of TSSs to search for TOP and TOP-like motifs.…”

“…We used tiling array data (Danks et al, 2013) and CAGE (Danks et al, 2018) data generated from O. dioica oocytes to define oocyte-stocked mRNA transcripts.…”