Paused Pol II coordinates tissue morphogenesis in the Drosophila embryo

Abstract

Paused RNA polymerase (Pol II) is a pervasive feature of Drosophila embryos and mammalian stem cells, but its role in development is uncertain. Here, we demonstrate that a spectrum of paused Pol II determines the "time to synchrony"-the time required to achieve coordinated gene expression across the cells of a tissue. To determine whether synchronous patterns of gene activation are significant in development, we manipulated the timing of snail expression, which controls the coordinated invagination of ∼1,000 mesoderm cells during gastrulation. Replacement of the strongly paused snail promoter with moderately paused or nonpaused promoters causes stochastic activation of snail expression and increased variability of mesoderm invagination. Computational modeling of the dorsal-ventral patterning network recapitulates these variable and bistable gastrulation profiles and emphasizes the importance of timing of gene activation in development. We conclude that paused Pol II and transcriptional synchrony are essential for coordinating cell behavior during morphogenesis.

Figure 1. BMP/Dpp target genes exhibit distinct coordination profiles

(A-H) cc14 embryos hybridized with tup and pnr fluorescent (magenta) intronic probes for detecting nascent transcripts (nuclei stained with DAPI (blue)). Raw images for tup and pnr transcripts are shown in B, D, and the corresponding processed images are shown in B′ and D′. Images shown in B and D are magnifications of bracketed regions in A and C. (E-H) tup (E,G) and pnr (F,H) expression during mid (E,F) and late (G,H) cc14. (I,J) Dynamics of gene expression during cc14 based on the fraction of nuclei containing nascent transcripts. (I) Endogenous tup expression (blue) reaches 50% of the complete pattern (t50=26) 15min earlier than does pnr (black) (t50=41). (J) There is a delay in tup dynamics when the minimal promoter of a tup BAC transgene (tupY) is replaced by that of pnr (tupY-PnrPr) (see also Figure S2). The red curves represent the fitted curves (using a cumulative gamma distribution) to the data depicted in panels I and J (see supplementary information, Figure S1 and Table S1). T50 values are determined from these fitted curves.

Figure 2. The minimal promoter mediates paused Pol II

(A) Pol II Chip-Seq reads of the pnr/tup transgene in a tissue where it is silent. (B) Pol II ChIP followed by qPCR showing enrichment at the tupPr/yellow junction. Error bars represent SD. (C) Permanganate footprinting reveals a promoter-proximal “transcription bubble” in mutant embryos where the tupPr/yellow transgene is silent. (D) Reduced levels of maternal Trl (turquoise) or NelfE/Spt5 (pink) cause a delay in the expression profile of the pnrE>tupPr transgene. A similar effect is observed with a truncated version of the tup promoter lacking the upstream GAGA sites.

Figure 3. A spectrum of synchrony

The pnr intronic enhancer (PnrE) was placed upstream of the tup, sna, Hsp70, sog, ths and pnr promoters (Pr) (see diagram in upper left). (A-D) Examples of transgenic embryos stained with a yellow intronic probe at the midpoint of cc14. The tupPr mediates synchronous expression in the dorsal ectoderm (A), while the pnrPr mediates stochastic expression (D). (E) Temporal coordination profiles during cc14. The tup promoter provides the rapid coordination profile, while the pnr promoter exhibits the slowest coordination. Sog and ths give intermediate. (F) Relative amounts of Pol II at the promoter regions of inactive genes. For actively expressed genes, we denote them as “expressed”, the normalized Pol II reads are provided in TableS2.

Figure 4. Minimal promoters are sufficient to perturb snail temporal coordination

The distal snail enhancer (snaE) was placed upstream of the snaPr (A), sogPr (B), and thsPr (C) promoters attached to the yellow reporter gene (see diagram in upper left). (A-C) Processed images after fluorescent in situ hybridization using a yellow intronic probe. (D) Temporal coordination profiles during cc14. (E) High resolution confocal image of yellow mRNAs encoded by the snaE>snaPr/yellow minigene. Arrowheads: individual cytoplasmic mRNAs; arrow: nascent transcripts. (F) Bar graph showing the estimated promoter strenght from the pnr, ths and sna promoters just prior to gastrulation (see Experimental Procedures and Figure S4). Error bars represent SD.

Figure 5. Stochastic expression of snail results in gastrulation defects

Transgenic rescue embryos stained with a snail (sna) probe (in red) at gastrulation stages (A-C, E-G, I-K) and correspondant invaginating cells false colored in green (A′-C′, E′-G′, I′-K′) . (A-C) When sna expression is driven by a snaBAC-sna promoter lacking the primary enhancer, all embryos gastrulate normally (A, C′). Variable gastrulation defects are obtained when the sna promoter is replaced by the moderately paused sog promoter (E-G). Most embryos show pockets of ingressing cells (E, E′) and “half furrow” (F, F′) and occasional embryos show a normal furrow (G, G′). When sna is artificially depaused by replacing its promoter by the ths promoter sequence (I-K), most embryos fail to gastrulate (I-J′) but rare embryos exibit an extended groove of invaginating cells (K, K′). (D, H, L) Transgenic embryos stained with snail (red) and twist (green) antibodies at gastrulation when the ventral furrow is invaginating. See also Figure S5.

Figure 6. Modeling gastrulation variability: the importance of coordination

(A) Mesodermal region of a DAPI stained embryo to show the segmentation process of the nuclei. The panel below is a schematic illustrating the neighbors (j) of a given mesodermal nuclei (i). We allow for nearest neighbor diffusion where i′th nucleus is diffusive coupled to its nearest neighbors that share a boundary (j=1:6 in this case). (B) Simplified mathematical model for Snail dynamic expression in a given nucleus (i). The key parameters are: the timing of snail activation in the particular nuclei, the concentration of the neurogenic repressor (Rep) and the concentration of activators like Dorsal (k1), number of nearest neighbors (NN) and the strength of the diffusive coupling between nuclei (D). (C) Activation curves computationaly obtained for three different promoters, sna, sog and ths. (D-F) Results of computational simulations when snail temporal coordination is affected, t50 values are indicated. See also Figures S5, S6 and S7.