Replication fork reversal after replication-transcription collision

Abstract

Replication fork arrest is a recognized source of genetic instability, and transcription is one of the most prominent causes of replication impediment. We analyze here the requirement for recombination proteins in Escherichia coli when replication-transcription head-on collisions are induced at a specific site by the inversion of a highly expressed ribosomal operon (rrn). RecBC is the only recombination protein required for cell viability under these conditions of increased replication-transcription collisions. In its absence, fork breakage occurs at the site of collision, and the resulting linear DNA is not repaired and is slowly degraded by the RecJ exonuclease. Lethal fork breakage is also observed in cells that lack RecA and RecD, i.e. when both homologous recombination and the potent exonuclease V activity of the RecBCD complex are inactivated, with a slow degradation of the resulting linear DNA by the combined action of the RecBC helicase and the RecJ exonuclease. The sizes of the major linear fragments indicate that DNA degradation is slowed down by the encounter with another rrn operon. The amount of linear DNA decreases nearly two-fold when the Holliday junction resolvase RuvABC is inactivated in recB, as well as in recA recD mutants, indicating that part of the linear DNA is formed by resolution of a Holliday junction. Our results suggest that replication fork reversal occurs after replication-transcription head-on collision, and we propose that it promotes the action of the accessory replicative helicases that dislodge the obstacle.

Figure 1. Replication fork reversal model and schematic representation of the I-SceI fragment carrying the inverted region in InvA and InvBE.

A Replication fork reversal model. In the first step (A), the replication fork is arrested, and the leading and lagging strand ends of the newly synthesized strands anneal. The reversed fork forms a four-arm structure (Holliday junction, HJ; two alternative representations of this structure are shown, open X and parallel stacked X). RecBC is essential for resetting of the fork, either by RecA-dependent homologous recombination (B–C) or by DNA degradation (B–D). Either pathway creates a substrate for replication restart proteins (PriA and its partners). In the absence of RecBCD (E), resolution of the HJ causes chromosome linearization. Continuous lines: parental chromosome. Dashed lines: newly-synthesized strands. Circle: RuvAB. Incised circle: RecBCD. B Schematic representation of the I-SceI fragment carrying the inverted region in InvA and InvBE. Positions of inversion ends (att) are indicated by wide flat arrowheads. Positions of the cleavage sites (I-SceI, NotI) are indicated by vertical arrows. Grey boxes represent the rrn genes and the open circle shows the position of oriC. The direction of replication and the direction of transcription of the inverted rrn are indicated. The distances between the two I-SceI sites and between the I-SceI sites and the 3′ end of the rrn operons are shown. The drawing is not to scale.

Figure 2. Inv recB strains are LB–sensitive.

Appropriate dilutions of overnight cultures grown at 37°C in MM (OD 0.8 to 1.5) were plated on MM and LB plates, which were incubated at 37°C. White boxes: colony forming units (cfu)/ml on MM plates after 48 h incubation; grey boxes: cfu/ml on LB plates after 16–24 h incubation. Bars indicate standard deviations. Top: InvBE strains; bottom: InvA strains. ruv stands for ruvAB inactivation. rpo* stands for the rpoC ^Δ215–220 mutation, pR stands for pEM001, the plasmid encoding RNaseH. Colonies were small in 48 h on MM for Inv recB ruv and Inv recB ruv recG mutants, but a similar growth delay was observed for non-inverted strains. A small percentage of small colonies was observed after two to four days of incubation on LB with the InvBE recB ruvAB (or recG) and InvBE recB ruvAB recG mutants, however the number of these colonies was highly variable and no colony was ever observed with the InvA recB ruvAB (and/or recG) mutants, indicating that these mutants still require RecBC for growth on rich medium.

Figure 3. Chromosome breakage in InvA recB and InvA recA recD mutants.

Chromosomes of InvA recombination mutants, grown in minimal medium (MM) or for 1, 2 or 3 hours in LB, were cleaved with the I-SceI enzyme and fragments were separated by PFGE. A. Ethidium bromide stained gel with I-SceI cut chromosomes from InvA recB and InvA recB recJ mutants. First lane Saccharomyces cerevisiae chromosome ladder, relevant sizes are indicated on the left. B. Southern blot of the gel shown in A using the origin-proximal probe in the 800 kb I-SceI fragment, both the intact fragment and fragments of smaller sizes hybridize with the probe. C. Southern blot of the gel shown in A using the origin-distal probe in the 800 kb I-SceI fragment, only the intact I-SceI fragment hybridizes with the probe. D. Ethidium bromide stained gel with I-SceI cut chromosomes from the InvA recA recD mutant. E. Southern blot of the gel shown in D using the origin-proximal probe in the 800 kb I-SceI fragment, both the intact fragment and fragments of smaller sizes hybridize with the probe. F. Schematic representation of the different DNA fragments. The triangles represent the two rrn operons as indicated, the black circle represent oriC and the little bars above the lines represent the position of the origin-proximal (left) and origin distal (right) probes.

Figure 4. Chromosome breakage in InvBE recB and InvA recA recD mutants.

Chromosomes of the indicated InvBE mutants, grown in minimal medium (MM) or for 1, 2 or 3 hours in LB, were cleaved with the I-SceI enzyme and fragments were separated by PFGE. A. Ethidium bromide stained gel with I-SceI cut chromosomes from InvBE recB and InvBE recB recJ mutants. First lane Saccharomyces cerevisiae chromosome ladder, relevant sizes are indicated on the left. B. Southern blot of the gel shown in A using the origin-proximal probe in the 800 kb I-SceI fragment, both the intact fragment and fragments of smaller sizes hybridize with the probe. C. Southern blot of a gel made with the InvBE recB I-SceI cut chromosomes, using the origin-distal probe in the 800 kb I-SceI fragment, only the intact I-SceI fragment hybridizes with the probe. D. Ethidium bromide stained gel with I-SceI cut chromosomes from the InvBE recA recD mutant. E. Southern blot of the gel shown in D using the origin-proximal probe in the 800 kb I-SceI fragment, both the intact fragment and fragments of smaller sizes hybridize with the probe. F. Schematic representation of the different DNA fragments. The triangles represent the three rrn operons as indicated, the black circle represent oriC and the little bars above the lines represent the position of the origin-proximal (left) and origin-distal (right) probes.

Figure 5. rrn operons are a barrier to DNA degradation.

Top; schematic representation of the I-SceI fragment carrying the inverted rrnE operon. Triangles represent rrn operons as indicated, the black circle represents oriC, the black line represents the DNA I-SceI fragment and the numbered bars under this line show the positions of the different probes. Bottom; chromosomes of InvBE recA recD cells grown in MM or in LB for 2 hours were cleaved with I-SceI and fragments were separated by PFGE, for each panel: left lane, cells grown in MM, right lane, cells grown in LB for 2 hours. Southern blots were hybridized with the different probes indicated above each panel. From left to right: probe 1 - origin-proximal probe, probe 2 - rrnC promoter probe, probe 3 - rrnC terminator probe, probe 4 - rrnA promoter probe, probe 5 - rrnA terminator probe, probe 6 - origin-distal probe. A schematic representation of the fragments of different length is shown on the right. For each probe all hybridizing fragments are necessarily larger than the distance between the origin-proximal I-SceI site and the probe, so that the smear stops at the position of the probe.

Figure 6. Fork breakage is partially RuvABC-dependent in InvA mutants.

A. Chromosomes of the indicated InvA mutants, grown in minimal medium (MM) or for 1, 2 or 3 hours in LB, were restricted with the NotI enzyme and fragments were separated by PFGE. A. Ethidium bromide stained gel with NotI restricted chromosomes from InvA recB and InvA recB ruvAB mutants. Fragment sizes are indicated on the left. The star indicates the position of migration of the 171 kb DNA fragment formed by fork breakage and DNA degradation. B. Southern blot of the gel shown in A using the rrnC promoter probe, both the intact 208 kb NotI fragment and fragments of smaller sizes hybridize with the probe (the minor hybridization with the 193 kb Not1 restriction fragment may result from co-migration of broken DNA with this fragment). C. Southern blot made with the gel shown in A, using the rrnC terminator probe, only the intact 208 kb NotI fragment hybridizes with the probe. D. Schematic representation of the different DNA fragments. The triangles represent the two rrn operons as indicated, the black circle represents oriC and the bars above the lines show the positions of the probes. E. For each mutant, Southern hybridizations of 3 to 6 gels were quantified, and the percentage of hybridized DNA that remains in wells, that migrates at the 208 kb position and that migrates at the 171 kb position were calculated. For clarity, only the percentages of migrating DNA are shown here (see Table S2 for complete results); light grey, 208 kb fragment (intact), dark grey, 171 kb fragment (resulting from fork breakage and DNA degradation up to rrnC). Vertical bars indicate standard deviations.

Figure 7. Model for the restart of forks arrested by a highly expressed, oppositely oriented rrn operon.

A blocked fork is either reversed (RFR) or re-replicated by a following round of replication initiated at the replication origin (re-replication). The product of RuvABC-catalyzed resolution of the HJ formed by fork reversal, and the product of re-replication are similar origin-proximal dsDNA ends (left part of the model). These DNA ends are repaired in Rec⁺ cells by homologous recombination catalyzed by RecBCD, RecA and RuvABC (not shown), but remain unrepaired in recBC and recA recD mutants, where they are detected by electrophoresis of Not1- or I-Sce1-treated DNA (Figure 3, Figure 4, Figure 5, Figure 6). In Rec⁺ cells the dsDNA end formed by fork reversal can be directly acted upon by RecBCD (see Figure 1) and processed by either homologous recombination (RecBC(D)-RecA-RuvABC pathway) or by DNA degradation (RecBCD (exo V) pathway). Reversed forks resetting by either pathway produces a replication fork that has moved backward, further from the obstacle than the original blocked fork. We propose that the reloading of new replisome at such forks favors the binding of a second accessory helicase (DinG or UvrD), required with Rep for replication across the inverted rrn operon.