Fig 1.
Identification of differentially decaying operons in E. coli.
(A-C), Differential decay in three representative E. coli operons, depicted by normalized RNA-seq coverage in steady state (black, t = 0) or at two time points (green and red) following rifampicin treatment. RNA-seq coverage was normalized by the number of uniquely mapped reads in each library. Bar graphs show average half-life calculations from three replicates with error bars representing standard deviation. (D), Ratio of steady-state mRNA abundance (blue) and mRNA half-lives (red) shown for a subset of regulated gene-pairs in which mRNA abundances and decay rates closely matched. Gene names are marked below the x-axis. Shown is the average ratio between the genes with error bars denoting standard deviation.
Fig 2.
Protective RNA 3’ structures at the boundaries of differentially decaying transcript regions.
(A), A model for differential decay in operons in which the 5’-most gene is preferentially stabilized, as described for the maltose operon [8]. Scissors and Pac-Man represent endonucleases and 3’-5’ exonucleases, respectively. (B-D), Differential decay in three representative E. coli operons, depicted by normalized RNA-seq coverage in steady state (black, t = 0) or at two time points (green and red) following rifampicin treatment. Bar graphs show average half-life calculations with error bars representing standard deviation. RNA-seq coverage was normalized by the number of uniquely mapped reads in each library. RNA 3’ ends detected by term-seq are shown as black arrows with the height of the arrow representing the total number of supporting reads. The sequences present immediately upstream to the recorded 3’ ends were folded using RNAfold [35] and are shown to the right of each coverage plot, with the estimated structure stability, measured in kcal/mol.
Fig 3.
RNase E cleavage and protective 5’-end structures correlate with differential decay.
(A), A model for differential decay in operons in which the middle or 3’-most gene is preferentially stabilized, as described for papBA [12]. Scissors and Pac-Man represent endonucleases and 3’-5’ exonucleases, respectively. (B-D), Differential decay in three representative E. coli operons, depicted by normalized RNA-seq coverage in steady state (black, t = 0) or an additional time points (green) following rifampicin treatment. RNA-seq coverage was normalized by the number of uniquely mapped reads in each library. The position and number of reads supporting RNase E cleavage sites in the WT strain or in the RNase E mutant are shown as dark and light orange arrows, respectively. The height of the arrows represents the total number of supporting 5’ end reads across all three replicates, normalized by the number of uniquely mapped reads in each experiment. The predicted structure and stability of the RNA sequence present immediately downstream of the RNase E cleavage site end is shown next to each gene, with blue rectangles specifying the position of the structure in the genome. The 5’ cleavage position is marked by a scissors cartoon.
Fig 4.
Ribosome density can guide differential operon decay.
(A), Comparison of the relative ribosome densities across uniformly (n = 533) and differentially decaying operonic gene-pairs for which ribosome densities were previously measured (n = 39) [3] is shown as grey and green box-plots, respectively. Outliers are marked as red dots and the median is marked as a horizontal line within the box. The distributions were compared using a two-sided Wilcoxon rank-sum test (p < 10−8). On the left, a drawing illustrates the growing differences in ribosome densities between the genes along the boxplot’s y-axis. (B), Illustration of the effect of the translation-initiation inhibitor kasugamycin and its hypothesized effect on ribosome densities in polycistronic transcripts. (C), A scatter plot showing the change in relative decay rate of regulated gene-pairs calculated using recently published decay rates [22] for control (x-axis) and kasugamycin treated (y-axis) bacteria (S8 Table). The y = x function (denoting 1:1 ratio) is shown as a dashed line.
Fig 5.
Proposed models for regulation of differential decay rates within the same transcript.
(A), A model for stabilization of the upstream gene. The mature polycistronic transcript is protected from 3’-5’ exonucleases by the 3’ terminator hairpin structure. The relatively higher ribosome density over gene A provides protection from endonucleases, leading the RNase E to cleave the gene B segment, which results in the removal of the protective terminator structure, exposing gene B to rapid digestion by 3’-5’ exonucleases. Gene A remains protected by a 3’ RNA structure that blocks processive exonuclease activity. (B) A model for stabilization of the downstream gene. The mature operon is protected by the terminator structure. The relatively higher ribosome densities over gene B guide the initial cleavage to gene A. 3’-5’ exonucleases degrade gene A while gene B maintains its ribosome densities and protective terminator structure. Endonuclease cleavage occurs upstream of a protective 5’ structure that further protects the stable transcript from RNase E.