Perturbation of mRNP biogenesis reveals a dynamic landscape of the Rrp6-dependent surveillance machinery trafficking along the yeast genome

Abstract

Eukaryotic cells have evolved a nuclear quality control (QC) system to monitor the co-transcriptional mRNA processing and packaging reactions that lead to the formation of export-competent ribonucleoprotein particles (mRNPs). Aberrant mRNPs that fail to pass the QC steps are retained in the nucleus and eliminated by the exonuclease activity of Rrp6. It is still unclear how the surveillance system is precisely coordinated both physically and functionally with the transcription machinery to detect the faulty events that may arise at each step of transcript elongation and mRNP formation. To dissect the QC mechanism, we previously implemented a powerful assay based on global perturbation of mRNP biogenesis in yeast by the bacterial Rho helicase. By monitoring model genes, we have shown that the QC process is coordinated by Nrd1, a component of the NNS complex (Nrd1-Nab3-Sen1) involved in termination, processing and decay of ncRNAs which is recruited by the CTD of RNAP II. Here, we have extended our investigations by analyzing the QC behaviour over the whole yeast genome. We performed high-throughput RNA sequencing (RNA-seq) to survey a large collection of mRNPs whose biogenesis is affected by Rho action and which can be rescued upon Rrp6 depletion. This genome-wide perspective was extended by generating high-resolution binding landscapes (ChIP-seq) of QC components along the yeast chromosomes before and after perturbation of mRNP biogenesis. Our results show that perturbation of mRNP biogenesis redistributes the QC components over the genome with a significant hijacking of Nrd1 and Nab3 from genomic loci producing ncRNAs to Rho-affected protein-coding genes, triggering termination and processing defects of ncRNAs.

Keywords: ChIP-seq; RNA-seq; mRNP biogenesis; mRNP quality control; yeast.

Figure 1.

Fate of a subpopulation of mRNAs down-regulated by Rho action in Wt cells and their rescue in a strain with a deletion of RRP47, i.e. depleted of Rrp6. a) MAplot of the log2FoldChange ratios of the mRNAs between -Rho and + Rho conditions (log2(±Rho) from WT cells distinguishes three sub-populations: statistically upregulated in red, statistically downregulated in blue and no statistical variation in grey. The dashed blue circle surrounds the population of interest which is down-regulated. b) MAplot as in a but for the mRNAs obtained from rrp47Δ cells and focuses only on the subset of mRNAs that were down-regulated in WT cells as determined in a.

Figure 2.

Meta-transcript profiles variations of mRNAs obtained from WT and rrp47Δ cells in the absence or presence of Rho. The top left graph shows the average reads coverage for all analyzed mRNAs. The bottom left graph represents the down-regulated mRNAs subpopulation as determined in Figure 1A. The two right graphs display the two groups of the down-regulated mRNAs sorted according to their rescue in the rrp47Δ genetic background (Rescued) or not (Non-rescued) as determined in Figure 1B. For all the graphs, the dark blue line represents mRNAs obtained from WT cells in the absence of Rho whereas the light blue gives the meta-profiles of the mRNAs obtained from the same cells but in the presence of Rho. Light green and orange display the meta-profiles of the mRNAs obtained from the rrp47Δ strain in the absence and presence of Rho, respectively.

Figure 3.

Differential effect of Rho on mRNAs and ncRNAs. The Log2 ratios of RNA levels between +Rho and -Rho conditions are plotted on the metagenes for each RNA biotype as indicated on the top of each graph. Blue lines are for RNAs extracted from Wt cells and green lines are for RNAs obtained from the rrp47Δ strain. Standard error is represented by the lightened area around each line.

Figure 4.

Rho activity in yeast affects the termination and processing of ncRNAs. a) IGV snapshots of raw RNA-seq reads focusing on a snoRNA, a SUT and a CUT. Each snapshot shows the reads coverage of RNAs extracted from WT or rrp47Δ strains and with or without Rho induction. The scale for the number of reads obtained from the raw RNA-seq data is shown in brackets on the right of each snapshot. Red arrow for snR51 indicates the annotated termination site for the snoRNA. b) Scatter plot of Log2 normalized reads counts in a 100 bp region downstream of annotated genomic features for RNAs extracted from WT cells under Rho perturbation versus the same measure for RNAs obtained from unperturbed cells. Dashed line shows the cut-off of L2FC = 1. The color code for the analyzed RNA biotypes is inside the graph.

Figure 5.

Rho activity mediates the stimulation of massive recruitment of QC components to Rrp6-sensitive mRNAs genomic loci. a) Beeswarm plot of log2 fold enrichment detected on the rescued mRNAs genomic loci obtained by ChIP-seq experiments for each indicated QC protein. The ChIP signals are shown as open dots in blue for -Rho and in red for +Rho conditions. A student t-test was used to assess the significance of enrichment differences between the -Rho and +Rho conditions with the following code 0.05>*>0.01>**>0.001>***. (NA) stands for not applicable. The table above the graphs gives the number of analysed features (n_f) and the number of loci possessing at least one peak for each protein (n_p) in -Rho or +Rho conditions. b) Correlogram of Empirical Distribution Cumulative Function correlation area (calculated with the genometricorr package) for each protein pairs over the analysed genomic loci under Rho perturbation (“+“symbols) or not (‘-’ symbols). The crossed boxes indicate the absence of correlation significance. c) IGV snapshots of ChIP-seq peaks mapped over two rescued mRNAs genomic loci. The ChIP signals for each protein are shown both for -Rho and +Rho conditions.

Figure 6.

Eviction of QC components from ncRNAs genomic features under conditions of perturbation of mRNPs biogenesis by Rho. Beeswarm plots of log2 fold enrichment obtained by ChIP-seq experiments for each indicated QC protein detected on CUTs and SUTs in (A) and on snoRNAs genomic loci in (B). The symbols and labelling are the same as in Figure 5A. (NS) stands for non-significant according to the Student t-test which is related to the difference in the intensity of the peaks. (NA) stands for not applicable owing to the lack of a significant number of peaks in the -Rho condition.

Figure 7.

Summary of the observed dynamic landscape of QC components over the yeast genome. Circos plots summarising the ChIP-seq results for the four QC proteins over the 16 yeast chromosomes analysed under -Rho (left plots) and +Rho (right plots) conditions. The two top plots show the ChIP signals detected for the ncRNAs genomic loci (CUTs, SUTs and snoRNAs). The two plots at the bottom show the ChIP signals detected for the genomic features of the Rho-affected mRNAs that were rescued by Rrp6 depletion. The red arrow symbolizes the hijacking of Nrd1 and Nab3 from ncRNAs genomic loci to the Rho-affected mRNA genes upon perturbation of mRNP biogenesis by Rho.