RNA polymerase stalling at developmental control genes in the Drosophila melanogaster embryo

Abstract

It is widely assumed that the key rate-limiting step in gene activation is the recruitment of RNA polymerase II (Pol II) to the core promoter. Although there are well-documented examples in which Pol II is recruited to a gene but stalls, a general role for Pol II stalling in development has not been established. We have carried out comprehensive Pol II chromatin immunoprecipitation microarray (ChIP-chip) assays in Drosophila embryos and identified three distinct Pol II binding behaviors: active (uniform binding across the entire transcription unit), no binding, and stalled (binding at the transcription start site). The notable feature of the approximately 10% genes that are stalled is that they are highly enriched for developmental control genes, which are either repressed or poised for activation during later stages of embryogenesis. We propose that Pol II stalling facilitates rapid temporal and spatial changes in gene activity during development.

Fig. 1. Different classes of Pol II binding profiles

ChIP-chip assays were performed with 2–4 hr Toll^10b embryos using antibodies that recognize both the initiating and elongating form of Pol II. The enrichment ratios of Pol II are shown on the y-axis. (A–D) display the binding patterns across genes that are repressed in Toll^10b embryos. All four genes display high levels of Pol II near the transcription start sites. At some genes such as tup (A), Pol II is tightly restricted to this region, whereas at other genes including sog (C) and brk (D), Pol II is also detected at lower levels throughout the transcription unit. (E,F) Pol II is uniformly distributed across the transcription units of genes that are actively transcribed. The Heartbroken (Hbr, also called stumps or Dof—downstream of FGF) locus (E) is specifically activated in mesodermal precursor cells, while RpL3 (F) is a ribosomal gene. (G,H) No Pol II binding is found at many genes that are inactive during embryogenesis. The eyeless (ey) gene (G) is expressed during eye development at larval stages but not in the early embryo. Likewise, the torso (tor) gene (H) is only active during oogenesis but not in the early embryo.

Fig. 2. Whole-genome analysis of Pol II binding

(A) Genes were assigned to either one of three classes - stalled Pol II, active Pol II or no Pol II, and - based on their Stalling Index (box). The Stalling Index is the ratio between the maximum enrichment near the transcription start site (max_TSS) (± 300 bp) and the median enrichment of the probes distributed across the transcription unit (median_transcript) (excluding the first 600 bp). Stalling Index values of >4 qualified as “stalled Pol II”, whereas Stalling Index values <2 qualified as “uniform (active) Pol II”. If no probe within the TSS region was significantly bound, the gene was assigned to the “no Pol II” category. (B) Over 76% of all protein-coding genes could be assigned to one of three categories based on Stalling Index values: 12% have stalled Pol II, 23% display the active form of Pol II and 37% of genes have no Pol II. Among the genes with stalled Pol II, 62% have Pol II tightly restricted to the transcription start site.

Fig. 3. Confirmation of the class of genes with Pol II stalling

(A) The group of stalled genes contains all known heat shock genes where Pol II stalling has been well documented. (B) Metagene analysis shows that the average peak of “stalled Pol” II is ~ 50 bp downstream of transcription (arrow). The profile for “no Pol II” and “active Pol II” is shown as comparison. (C) Analysis of the transcript levels confirms that genes with the stalled Pol II profile (tightly restricted to the transcription site) are either silent or expressed at low levels. Genes that show Pol II enrichment at comparable levels throughout the transcription unit (active Pol II) are expressed at significantly higher levels. Genes with no Pol II are not expressed. The transcript levels are represented as box and whiskers plot of the fold-ratios (measured by whole-genome tiling arrays20). The box represents the 25^th and 75^th percentile with the median as red bar. The whiskers refer to the first and 99^th percentile. The scale on the y-axis is a log scale. (D) A permanganate footprint assay of the rho gene confirms stalled Pol II downstream of the transcription start site. Genomic sequences of A+G are shown as marker (lane 1). In comparison with purified genomic DNA, which was either not treated (lane 2) or treated (lane 3) with KMnO₄, a prominent hypersensitive T residue is detected in Toll^10b mutant embryos (lane 4), implying the existence of a transcription bubble at the region around +37 in vivo. The bottom panel shows actual sequences from +26 to +47 of the rho locus (relative to TTS as +1).

Fig. 4. Functional analysis of the three classes of genes

Representative categories of enrichment among genes with stalled Pol II, active Pol II or no Pol II for gene sets are shown for the (A) ImaGO database, which contains the in situ expression patterns of a substantial fraction of all protein coding genes in the Drosophila genome and (B) the gene sets from the Biological Process categories in the Gene Ontology (GO) database. The scale bars below indicate the significance of each test (hypergeometric distribution). Genes containing stalled Pol II (green) are significantly enriched for genes expressed in a subset of cells and in those of developing ectoderm (AISN = anlage in statu nascendi) at the time of the analysis (120–240 min). The stage of the enriched categories is indicated in parenthesis (s1–3 = ~0–100 min, s4–6 = ~100–200 min, s7–8 = ~200–250 min). The genes expressed as subset (s4–6) are largely identical to those of the ectoderm AISN (s4–6) category because mesoderm, neurectoderm and dorsal ectoderm are specified at that stage. Functional analysis confirms that genes with stalled Pol II are enriched for genes with roles in development, in particular those required for neurogenesis, ectoderm differentiation, and muscle development. Genes displaying active Pol II (red) are enriched for genes that display ubiquitous expression in developing embryos. They are enriched for functions that mediate cell proliferation and metabolic functions such as protein and nucleotide metabolism. Genes lacking Pol II (grey) tend to be inactive during embryogenesis, and deployed at later stages of the life cycle such as for cuticle function and vision.

Fig. 5. Pol II profile comparison of genes in the repressed vs. active state

Genes are shown that are repressed in Toll^10b embryos but are active in either Toll^rm9/Toll^rm10 embryos (A, B) or gd⁷ embryos (C,D) at 2–4 h after fertilization, were examined by Pol II ChIP assays. The enrichment ratios of Pol II are shown on the y-axis: active state (red) and repressed state (blue). The results show that the degree of Pol II stalling is dependent on the gene’s activity, with some genes showing a complete switch between the stalled and active form of Pol II.

Fig. 6. Pol II stalling at genes prior to activation

In vivo permanganate footprints were performed on the genes Drop (Dr), ladybird (lbe) and tail-up (tup) using early wild-type embryos (2–4 h) and S2 cells (107 cells), a cell line derived from older embryos. Positions of A+G (lane 1) and the transcription start site (TSS, arrow) are shown for orientation. T residues sensitive to KMnO₄ treatment are shown for naked DNA as control (lane 2), S2 cells (lane 3) and early embryos (lane 4). Hypersensitive T residues in the early embryo sample (selected positions marked on the right) indicate an open transcription bubble. The footprints found at the muscle regulating genes Dr and lbe are similar to the one of tup, a gene known to be repressed at that stage. These results show that Pol II is found in a stalled form at the Dr and lbe genes prior to activation during embryonic development.