Novel structure may hold the key to DNA break repair
Super-resolution microscopy has allowed researchers in the department to visualise a process for the first time that appears to play a central role in DNA repair.
Snapshop analysis of E.coli before induction of double-strand break. RecA-GFP (green) is diffuse in most cells and spontaneous spots are distant from double strand ends (red)
Professor David Sherratt, postdoc Dr Christian Lesterlin and microscopy experts Drs Lothar Schermelleh and Graeme Ball have published their findings in Nature (1).
The research provides valuable insight into how a cell ensures that its DNA is continuously protected from the harmful effects of double-strand breaks, one of many types of DNA lesions that plague the cell.
Double-strand break (DSB) repair is common to all organisms, from bacteria to humans. One way in which this can occur is via homologous recombination, where the DNA is repaired through pairing of the broken strands with homologous sequences on sister loci. Conserved proteins, RecA or Rad51 in higher organisms, bind to the broken DNA ends and search for homology within the intact sister.
In some organisms such as bacteria, homologous recombination is the only mechanism available for DSB repair. If the sisters are close together as they are during DNA replication and shortly afterwards, then the broken strand can find the sister locus easily. But it is less clear what happens if the sister is a long way off.
‘Bacteria are different from eukaryotes because they don’t have extensive sister cohesion,’ explains Professor Sherratt. ‘So shortly after replication, the two sisters separate to opposite parts of the cell - a considerable distance of about 1.5-2 microns.’
He compares the problem to that of two people having a conversation. If they are sitting next to each other, they can talk without any assistance. But if they are in different towns, they will need to use a phone.
Although the sisters are apart the majority of the time in bacteria - about 70% of the cell cycle - genetic studies suggest that the cells still retain the ability to repair their DNA when needed. Professor Sherratt and colleagues wanted to explore how this could be happening.
Up until now, the group was unable to visualise repair beyond the fuzzy images obtained from conventional microscopy. But with the significant resolution improvement of three-dimensional structured illumination microscopy (3D-SIM) in the Deltavision-OMX Blaze system - part of the Micron Oxford Advanced Bioimaging Unit within the department - it is now able to image bacteria with unparalleled resolution in 3D and with multiple colours.
‘We’ve profited from recent advancements of the SIM technology that enabled us to extend the method to live cells,’ explains Dr Schermelleh, a group leader at Micron. ‘Up to now, most super-resolution imaging has been done on fixed cells, but we’ve managed to adopt the Blaze technology and apply it in a relatively short time to a biologically relevant process in living cells.’
Working with Dr Lesterlin, Micron researchers Drs Schermelleh and Ball optimised the microscope setup and tailor-made analytical tools to acquire and evaluate very fast time-lapse of bacterial cells. They used this first to visualise whether cells could undergo repair of an introduced DSB when the sister loci were no longer in close association.
‘With our approach we showed that throughout the cell cycle, cells can undertake efficient pairing and subsequent repair between the cut locus and the uncut homologous sister,’ says Dr Lesterlin. ‘Segregated sisters more than 1.5 micron apart can reverse the process of chromosome segregation by pairing the homologous sequences - a prerequisite to efficient repair. During this process, cell growth and division are inhibited until the repair is complete.’
The group then looked at the involvement of RecA in this process. They made a DSB and tracked the movement of green fluorescent protein-tagged RecA filaments, taking 3D-SIM images of the cells every 2 seconds over a period of 200 seconds so that a 4D (3D + time) super-resolution dataset could be computationally reconstructed.
This player requires Flash or an HTML 5 capable browser
Movie showing 3D structured-illumination-microscopy of a cell with a DSB-induced RecA-GFP bundle. Note the dynamic behaviour and architecture of the bundle with its central body and thin extensions
What they observed was a structure that had never been seen before – a very fine RecA filament, extending from a thicker central bundle, which seemed to be visiting a large proportion of the cell compartment.
When they followed the formation of the filament in real-time with wide-field deconvolution microscopy, they found that the thicker bundle itself grew from the cytoplasmic pool of RecA recruited to the cut locus after the break. Multicolour 3D-SIM revealed that the bundle was localised to the space between the membrane and the nucleoid.
Professor Sherratt and colleagues believe that this new and unexpected structure may play a central role in the search for homology when the sisters are apart within the cell.
‘The cut locus appears to move towards its sister using the RecA bundle as a substrate for the movement,’ he says. ‘Although we don’t understand how it works yet at the molecular level, the bundle may act like a microtubule-like cytoskeletal element that facilitates pairing. This could be either by a treadmilling process involving the organised turnover of subunits or by providing the track on which a motor transports the cut ends.’
To gather quantitative details about the bundles, the group carried out what Professor Sherratt calls ‘in vivo biochemistry’. ‘We wanted to put numbers on the process of assembly and disassembly of the bundle as it happens in the live cell,’ explains Dr Lesterlin. ‘Using this approach, we were able to generate a diagram which summarises the behaviour of the damaged region and RecA throughout the repair process.’
Another question was whether increased expression of RecA is needed for bundle formation. Although there is known to be a cytoplasmic pool of RecA and RecA overexpression follows after induction of DNA damage, the group found to its surprise that RecA expression is not required.
‘It’s an immediately available resource,’ says Dr Lesterlin. ‘In less than 5 minutes after induction of damage, the cytoplasmic supply of RecA forms into bundles. That went somewhat against what had been suspected before.’
The group also investigated the rate of turnover of RecA in the bundle structure using fluorescence recovery after photobleaching (FRAP). They found that the bundle is stable once formed and there is very little replacement of RecA monomers within it. After the bundle has carried out its function, it appears to be disassembled rapidly.
Schematic of DSB-end and RecA dynamics during DSB repair by E.coli homologous recombination, based on integration of all data
All the evidence from the extensive microscopy that the group carried out suggested that bundle formation is closely linked to the process of DNA repair. To establish an even stronger correlation, the group used RecA mutants that are unable to carry out DNA repair. As expected, they found that these mutants are also unable to make bundles.
Together, their results provide compelling evidence for a process in bacteria that facilitates the search for homology and subsequent DNA repair when the sisters are at a distance from each other. The process could well extend to the sorts of distances found in higher eukaryotes where the proteins and steps of homologous recombination are conserved.
Whilst the underlying mechanism remains a mystery, Dr Lesterlin comments that the group is in a strong position to start addressing this. ‘It’s important that we are studying a process [DNA repair] that has already been characterised genetically and biochemically very extensively,’ he says. ‘We know the different factors that are involved in each step, so we can look at these and work out what their role is in formation of bundles, the subsequent pairing and the completion of repair.’