Replication-Transcription Conflicts Generate R-Loops that Orchestrate Bacterial Stress Survival and Pathogenesis

Abstract

Replication-transcription collisions shape genomes, influence evolution, and promote genetic diseases. Although unclear why, head-on transcription (lagging strand genes) is especially disruptive to replication and promotes genomic instability. Here, we find that head-on collisions promote R-loop formation in Bacillus subtilis. We show that pervasive R-loop formation at head-on collision regions completely blocks replication, elevates mutagenesis, and inhibits gene expression. Accordingly, the activity of the R-loop processing enzyme RNase HIII at collision regions is crucial for stress survival in B. subtilis, as many stress response genes are head-on to replication. Remarkably, without RNase HIII, the ability of the intracellular pathogen Listeria monocytogenes to infect and replicate in hosts is weakened significantly, most likely because many virulence genes are head-on to replication. We conclude that the detrimental effects of head-on collisions stem primarily from excessive R-loop formation and that the resolution of these structures is critical for bacterial stress survival and pathogenesis.

Keywords: DNA replication; Listeria; R-loops; RNase H; accelerated evolution; gene orientation; pathogenesis; replication restart; replication-transcription conflicts; stress response.

Figure 1. DNA:RNA Hybrids Are Prevalent at Head-On Genes, Related to Figure S1

(A) Schematic of reporter constructs in either the co-directional (CD) or head-on (HO) orientation. spec^R, spectinomycin resistance gene. (B and C) DRIP-qPCR analysis of either wild-type (WT) cells or cells lacking rnhC carrying the lacZ reporters (B) or the luxABCDE reporters (C). CD, co-directional (black bars); HO, head-on (gray bars). Values are the means and SEM of at least two independent experiments.

Figure 2. Stable R-Loops that Form at Head-On Genes Block Replication

(A and B) ChIP-qPCR analysis of either WT or ΔrnhC cells carrying the lacZ reporters (A) or luxABCDE reporters (B). CD, co-directional (black bars); HO, head-on (gray bars). Data represent the means and SEM of at least three independent experiments. (C and D) 2D gel electrophoresis analysis cells with the head-on reporter gene lacZ (C) or luxABCDE (D). Top: DNA isolated from wild-type cells; middle: ΔrnhC; bottom: cartoon of the digested fragment and where the corresponding probe is located (red line). (E and F) Marker-frequency analysis of either WT or ΔrnhC cells carrying the lacZ reporter (E) or the luxABCDE reporter (F). The x axis indicates chromosomal location, and the y axis is the abundance of reads relative to the total number of reads in the sequencing library. Right: 42 kb portion of the genome surrounding the reporter constructs. The colored bars on the x axis indicate the reporter genes.

Figure 3. Unresolved R-Loops at Head-On Genes Cause Cell Death, Related to Figures S2 and S3

(A and B) Survival assays for either WT, ΔrnhC, or ΔrnhC cells with rnhC expressed at an ectopic locus (ΔrnhC rnhC⁺) carrying the reporter gene lacZ (A) or luxABCDE (B). Top: Representative plates; bottom: quantification. CD, co-directional (black bars); HO, head-on (gray bars). Values are the means and SD in colony forming units per mL (CFU/mL) of at least four independent experiments. (C and D) Survival assays for either WT cells or ΔrnhC cells carrying the reporter gene lacZ (C) or luxABCDE (D). Top: Representative plates; bottom: quantification. CD, co-directional (black bars); HO, head-on (gray bars). Values are the means and SD in colony forming units per mL (CFU/mL) of at least four independent experiments.

Figure 4. Overexpression of rnhC Reduces Mutation Rates of Genes in the Head-On Orientation

(A) Mutation rates calculated from measuring the reversion of the hisC952 allele in either WT cells or cells overexpressing rnhC (OE rnhC). CD, co-directional (black bars); HO, head-on (gray bars). The y axis indicates the mutation rate (mutations per 10⁸ cells per generation) as calculated from the Sandri-Sarkar Maximum Likelihood Estimator Method. Bars indicate the rate and 95% confidence interval calculated from at least 47 independent replicates. (B) Fluctuation analyses of data in (A) by the Luria-Delbrück method (described in Foster 2006), where P_r is the probability of at least r mutants and r is the number of mutants. Slope of −1 indicates fluctuation.

Figure 5. R-Loops Disrupt Head-On Gene Expression, Related to Figure S4

(A–C) qRT-PCR analysis from log phase cultures of either WT, ΔrnhC, ΔrnhC cells with rnhC expressed at an ectopic locus (ΔrnhC rnhC⁺) carrying the reporter gene lacZ (A) or luxABCDE (B). (C) Cells were induced and harvested in stationary phase. CD, co-directional (black bars); HO, head-on (gray bars). Data represent the means and SEM of at least three independent experiments. n.s., not significant. *p < 0.05.

Figure 6. Resolution of R-Loops at Head-On Genes Is Required for Survival of Stress, Related to Figure S5

(A) Global transcriptional analysis of genes upregulated during osmotic stress. CD, co-directional (black bars); HO, head-on (gray bars). Values are means and SEM. *p < 0.05. (B) qRT-PCR analysis of head-on osmotic stress-response genes in either WT (black bars) or ΔrnhC (gray bars) cells. The y axis indicates the fold induction of the indicated gene. Values are means and SEM of at least two independent experiments. (C) Representative plot of global analysis of DRIP-seq results in either WT or ΔrnhC cells during salt stress. Data represent the average normalized signal across all salt-response genes (including 100 bp upstream, 5′ UTR and 150 bp downstream, 3′ UTR) in each category (head-on and co-directional). Lines indicate the mean, and the shaded region indicates the standard error. (D) Representative plates from chronic survival assays of either WT or ΔrnhC cells exposed to stressors. (E) Quantification of WT (black bars) or ΔrnhC cells (gray bars) acutely treated with stressors. Values are the mean and SD of at least two independent experiments.

Figure 7. The Intracellular Pathogen Listeria monocytogenes Requires RNase HIII for Stress Tolerance and Pathogenesis, Related to Figure S6

(A) Representative plates from WT or ΔrnhC L. monocytogenes from chronic salt-exposure survival assays. (B) Growth curve of iBMDM cells infected with WT or ΔrnhC L. monocytogenes. Values are the mean and SD of three biological replicates. (C) Plaque assays from TIB73 hepatocytes infected with WT and ΔrnhC L. monocytogenes. n > 30 for each strain. Data are representative of two independent experiments. Values are the mean SD of biological replicates. (D) Survival assays from mice infected with WT or ΔrnhC L. monocytogenes. Values are CFU/organ. Solid lines indicate the median. hpi, hours post infection. *p < 0.05, **p < 0.01, ***p < 0.001.