Current Research
Genetic recombination is a fundamental cellular process and a major force in evolution. We study the mechanics of recombination and investigate the various pathways used to shuffle genes within and between species, to shape and reshape genomes and to repair damage. We are especially interested in how the interplay between recombination and DNA repair underpins chromosome replication and maintains genomic integrity. The age-related onset of cancer provides a stark reminder that no living organism can avoid damage to the genetic blueprint on which life depends. Damage is extensive and chronic. Organisms survive and reproduce because they normally repair such damage efficiently and use surveillance systems to make sure cells have completed all essential processes before they divide. But low levels of mutation and genomic rearrangement are generally tolerated, which emphasises the fact that evolution is concerned with survival rather than with exact transmission of the genome.
Our studies of recombination are centred on bacterial systems; we apply molecular, biochemical and structural methods to study key recombination proteins in Escherichia coli , and combine these with genetics and cell biology to investigate their function.
One major project focuses on RecG , a member of superfamily 2 of DNA and RNA helicases. We identified RecG initially as a protein that catalyses branch migration of the Holliday junction intermediate central to recombination, but we now have reason to believe it also underpins replication of damaged DNA. Although organisms generally have multiple systems for repairing DNA, some lesions inevitably slip the net. The dangers posed by unrepaired lesions lie in their ability to block advance of RNA and DNA polymerases, inhibiting transcription of damaged genes and preventing duplication of damaged chromosomes. Stalled transcription complexes present major obstacles to replication and also shield the underlying lesions. We discovered that RecG can convert a replication fork to four-way branched Holliday junction structure and have suggested ways in which this activity may facilitate rescue of replication forks stalled on the DNA template.
This movie shows the structure of RecG in a complex with a partial replication fork. RecG has conserved helicase domains (light blue) linked to a novel "wedge" domain (dark blue) providing specificity for binding branched DNA, with narrow channels accommodating ssDNA at the branch point.
A deep cleft between the helicase and wedge domains provides a channel for dsDNA. It also harbours a helical hairpin structure linked to an adjacent loop (green) projecting into the channel.
The adjacent movie shows the structure of RecG in a complex with a partial replication fork. RecG has conserved helicase domains (light blue) linked to a novel "wedge" domain (dark blue) providing specificity for binding branched DNA, with narrow channels accommodating ssDNA at the branch point. A deep cleft between the helicase and wedge domains provides a channel for dsDNA. It also harbours a helical hairpin structure linked to an adjacent loop (green) projecting into the channel.
A second movie shows further details of these structures in a section of RecG removing the wedge and exposing the dsDNA channel. The helical hairpin places two arginines (brown) in opposing positions where they are stabilised by a network of Hydrogen bonds involving a glutamate from helicase motif VI (red). We believe that disruption of this network, triggered by ATP hydrolysis, moves the adjacent loop (green) in the dsDNA binding channel and that a swinging arm motion of this loop drives DNA translocation, possibly via a ratchet mechanism involving contacts between a conserved glutamate (yellow) and the DNA backbone ( Mahdi et al 2003 ).
A second movie shows further details of these structures in a section of RecG removing the wedge and exposing the dsDNA channel. The helical hairpin places two arginines (brown) in opposing positions where they are stabilised by a network of Hydrogen bonds involving a glutamate from helicase motif VI (red).
We believe that disruption of this network, triggered by ATP hydrolysis, moves the adjacent loop (green) in the dsDNA binding channel and that a swinging arm motion of this loop drives DNA translocation, possibly via a ratchet mechanism involving contacts between a conserved glutamate (yellow) and the DNA backbone ( Mahdi et al 2003 ).
It has been proposed that translocation of dsDNA through the major cleft pulls the parental strands of a fork through the narrow channels flanking the wedge, neither of which is wide enough to accommodate duplex DNA, thus stripping off the nascent strands and allowing the parental strands to re-anneal. The unwound nascent strands may then also anneal, so that as the protein continues to translocate along the rewound parental duplex a "daughter duplex" is spooled out in front of the wedge ( Fig. 1 ).
Our current goals are to pin down the precise mechanism of DNA translocation by RecG and establish how strand separation and annealing are achieved at the wedge. These studies are expected to shed light on the motor activities of other proteins that translocate on dsDNA to catalyse strand cleavage (type I restriction enzymes), to dislodge stalled RNA polymerase molecules (Mfd protein), or to remodel chromatin (RSC/Sth1).
The idea that a damaged replication fork may be converted to a four-way branched molecule identical to the Holliday junction intermediate central to recombination holds many attractions. It allows recombination and repair enzymes to be brought into play, providing various means to remove or bypass the offending lesion, restore the fork and resume replication ( Fig. 2 ) ( Gregg et al 2002; McGlynn and Lloyd 2002 ).
Models of fork rescue emphasize the need for PriA to restart replication. This is consistent with the low viability and sensitivity to DNA damage of priA null cells. PriA initiates assembly of a primosome at fork and D loop structures via a series of protein-protein interactions culminating in transfer of the DnaB replicative helicase from a DnaB:DnaC complex to the PriA-DNA complex, allowing subsequent assembly of the DNA polymerase III holoenzyme.
The interaction of PriA with DnaB:DnaC is normally crucial. But certain substitutions in DnaC circumvent this requirement, enabling DnaB to be loaded without PriA and suppressing the phenotype of priA null strains. However, we found that a novel DNA binding protein, RdgC , is now required to maintain viability ( Moore et al 2003 ).
Our understanding of how recombination underpins replication in E. coli is complicated by the presence of two partially overlapping recombination pathways, one dependent on RecG helicase and the other on RuvABC Holliday junction resolvase . Both activities process Holliday junctions, but only RuvABC has the ability to cleave (resolve) these structures. The cleavage reaction is well documented and relies on the concerted action of the RuvAB branch migration and RuvC endonuclease activities ( Fig. 4 ).
E . coli encodes a second Holliday junction resolvase - the RusA protein - that can very effectively promote DNA repair and recombination in the absence of RuvABC, provided RecG is available. However, RusA is not normally expressed and can be deleted in ruv mutants without apparent effect. How junctions are resolved in the RecG pathway therefore remains a mystery - we are actively looking for the activity responsible.
In a collaboration with John Rafferty at the University of Sheffield ( Rafferty et al 2003 ) we have recently determined the crystal structure of the 14 kDa RusA protein and are currently trying to obtain crystals of RusA:DNA complexes. We have already identified the key residues essential for catalysis, but as with RuvC we have yet to determine how catalysis is targeted to specific DNA sequences. Understanding this reaction may reveal how an endonuclease has been engineered to become a dedicated Holliday junction resolvase.
Efficient catalysis by the RusA resolvase is limited to Holliday junction structures ( Bolt and Lloyd 2002 ). This specificity, and the ability to act alone, makes RusA a valuable tool for analysing junction processing in vivo . We exploit RusA to dissect the RuvABC and RecG pathways and investigate the crucial balance between DNA branch migration, which is necessary for interconversion of replication fork and Holliday junction structures, and DNA cleavage, which is needed to resolve junctions formed during genetic exchange, but which can also cause breakage of reversed forks ( Fig. 2 ).
Chromosomal DNA double-strand breaks are important lesions with potentially severe consequences for genome function and integrity. Our understanding of how breaks are repaired in E. coli has relied mostly on studies of cell survival following exposure to ionizing radiation and of recombination in genetic crosses presenting linear DNA substrates (conjugation and transduction).
We have recently established a new system in which we exploit the I- Sce I homing endonuclease to induce breaks in the E. coli chromosome ( Fig. 5 ). It provides a clear advantage over other systems in that a defined lesion is produced. Any factor required for survival can be assumed to be acting specifically in the repair of this lesion and no other. Furthermore, the break is induced at a specific location, enabling us to follow its formation and repair by physical means.
Using this system we have shown that as with UV repair and conjugational or transductional recombination, repair of a DNA double strand break can proceed via either of two recombination pathways. Both require RecBCD and RecA recombinases, and also PriA protein, presumably for replication of the missing DNA. But whereas one pathway depends on RuvABC resolvase the other does not and relies instead on RecG translocase. These findings are leading us to question whether the formation and subsequent resolution of classical Holliday junctions is essential for repair. The system we have established is opening new ways to investigate the RecG- and RuvABC- dependent recombination pathways. It also enables us to investigate the fidelity of repair, to probe the role of other factors implicated in double-strand break repair such as the SOS inducible RecN protein , and to investigate how repair is integrated with chromosome replication and cell division.
We are also interested in the effect transcription has on DNA recombination and repair. As stated above, stalled transcription complexes present major obstacles to replication and also shield the underlying lesions. In previous studies we discovered a striking correlation between the ability of E. coli cells to survive damage to DNA and their ability to modulate RNA polymerase ( McGlynn and Lloyd 2000 ). Specifically, we found that the stringent response regulators (p)ppGpp and certain RNA polymerase mutations that mimic the effect of (p)ppGpp increase the ability of strains lacking the RuvABC Holliday junction resolvase to survive irradiation with UV light . These RNA polymerase mutations introduce substitutions in regions of RpoB and RpoC flanking the DNA channel through the RNA polymerase complex ( Trautinger and Lloyd 2002 ). We have purified several of the mutant RNA polymerases and found that they reduce the stability of certain types of stalled RNAP complexes. We are currently investigating how this property improves the ability of ruv mutants to specifically withstand UV damage.
Acknowledgements
Our work is supported by programme and project grants from the Medical Research Council and The Wellcome Trust.
Research Collaborators
Dr Thorsten Allers (Royal Society University Research Fellow) Dr Ed Bolt (Wellcome Trust Career Development Fellow)
Lab Group Members
