By Matija Conic
Homologous recombination is one of the main pathways bacterial and eukaryotic cells utilise to repair double-stranded breaks in DNA (DSBs), the occurrence of which would otherwise lead to apoptosis or tumorigenesis. These can occur due to a variety of factors, including ionising radiation and indirect means such as arrested replication fork removal (either via fork regression or nuclease digestion) (Son et al, 2019). This review will focus on the molecular mechanisms through which this essential process takes place in both prokaryotes and eukaryotes.
Most common methods of DSB repair are non-homologous end joining (NHEJ) and homologous recombination (HR). HR in bacteria begins with an enzyme complex called RecBCD which associates with DSBs and starts to progressively degrade DNA into smaller fragments. RecC melts the two strands and feeds them into their respective helicases, RecB and RecD. In addition to helicase activity, RecB has a nuclease domain which degrades the 3’ end emerging from the motor domain. The 5’ end is also digested, albeit at a much lower frequency given that it’s less well positioned for digestion. This behaviour is altered once the 5’-GCTGGTGG-3’ cis-acting element (termed the crossover hotspot instigator, or chi site) is encountered by RecC (Watson et al, 1987). The chi site becomes tightly bound to RecC, while the 5’ end continues to be degraded, resulting in a long 3’ ssDNA overhang. This is critical for the next step of HR called strand invasion. RecBCD loads RecA proteins onto the resected ssDNA, producing a RecA nucleoprotein filament. These filaments can accommodate up to 4 ssDNA strands, though 1 and 3 strand intermediates are most common. Once the resected end is coated, RecA begins a homology search within the transformation-acquired dsDNA, which is transiently opened to check for regions of homology in a process involving base flipping. Once a 15 bp match is identified, a joint molecule is formed and strand invasion is initiated as the resected 3’ end base pairs with its complement within the neighbouring duplex (Watson et al, 1987).
This creates a D-loop formation containing a structure called a Holliday junction at the point where the strands cross. An analogous process occurs through the capture of the second DSB. The Holliday junctions are then translocated in a process called branch migration where base pairs are broken within the old duplex and created in the new heteroduplex. This occurs whilst the invading 3’-ends are extended through polymerisation to synthesise the new and replace the lost DNA. RuvA is responsible for junction recognition and the recruitment of two hexameric helicases called RuvB that hydrolyse ATP in order to power base pair exchange during branch migration (Watson et al, 1987).
RuvC is the last enzyme to associate, catalysing the resolution of Holliday junctions, the manner of which has a decisive effect on the recombination products. A Holliday junction can either be cleaved by cutting through the strands made up entirely of parental DNA or through cleaving the strands containing DNA of mixed origin. In the former scenario, the DNA from the 2 duplexes is “spliced” together, creating splice or crossover products characterised by the reassortment of the flanking genes. In the latter scenario, however, no reassortment occurs, with the flanking genes connected together by a patch of hybrid DNA, accordingly referred to as a “patch” or non-crossover product. If both Holliday junctions are cut in the same way, however, regardless of the cut type, a patch product will be created, while a splice product will ensue if the cuts are different. RuvC has a relatively common consensus sequence (5’-A/TTTG/C-3’) which roughly occurs every 64 base pairs, ensuring at least some branch migration occurs before resolution (Watson et al, 1987).
HR serves a multitude of purposes in bacteria. Given that chi sites are highly overexpressed in the E. Coli genome as opposed to any foreign phage DNA which may enter the cell, for instance, recombination is favoured over degradation, allowing the bacterium to effectively discriminate between the two and prevent potential damage.
Meiotic DSB induction occurs through a dimer called Spo11, a topoisomerase VI-like protein which is activated at specific moments of the cell cycle to cleave DNA. It does so through a transesterification reaction in a largely non-selective fashion, forming an energy-conserving 5’-phosphotyrosine linkage (Keeney et al, 2010).
Whilst bound to the 5’-end, Spo11 is removed by endonucleolytic cleavage, alongside a bound oligonucleotide. A single-stranded break is created by a complex consisting of Mre11, Rad50 and Xrs2 (in yeast) or Nbs1 (in mammals), hereafter referred to as MRX/N. Mre11 has 5’-3’ exonuclease activity, resecting the 5’ end 250-300 bp away from the DSB. Sae2/CtIP also associates with DNA and MRX/N and is thought to stimulate this activity, as well as contribute to the resection itself (Takeda et al, 2016). MRX/N localisation to DSBs occurs prior to the recruitment of RPA (the major eukaryotic ssDNA-binding protein) and Exo1, which binds DSBs with the help of RPA and MRX/N and performs more substantial 5’-3’ resection (over 10 kb). In addition to 5’-3’ exonuclease activity, Mre11 is also implicated in aiding the release of the Spo11-oligonucleotide adduct through 3’-5’ nucleolytic degradation (Symington, 2014). In order to avoid resection in G1 and M, Sae2/CtIP has CDK phosphorylation domains, rendering it active only in S and G2 when the sister chromatid templates necessary for HR are available (Heyer et al, 2010).
RPA efficiently prevents degradation and secondary structure formation on resected ends. Another key component, Rad52, binds Rad51, bridging it to RPA-bound ssDNA and exerting an inhibitory effect on RPA. RPA inhibition is the necessary precondition for Rad51-coating, which ensues after 2-5 Rad51 monomers initiate nucleation. A key tumour suppressor BRCA1 colocalizes to MRX/N and CtIP bound sites, contributing to 5’-3’ resection, in addition to recruiting Rad51 and BRCA2. BRCA2 in turn mediates the replacement of RPA by Rad51.
Similarly to RecA, the Rad51 filaments begin a homology search within an adjacent DNA duplex until a stretch of overlap is identified to allow for strand invasion. This is where the HR pathway diverges into two directions, depending on whether the second DSB end is engaged or not.
In double strand break repair, the second DSB end is engaged in one of two ways, either through another invasion event or through DNA annealing. Rad52 has a special role in this process, helping RPA-bound complementary stretches of ssDNA anneal to one another. The Holliday junctions are then translocated by ATP-powered helicases, including RecQ, RecG, FANCM, and Rad54, whilst the invading 3’-ends are extended through polymerisation (Heyer et al, 2010).
In synthesis-dependent strand annealing (SDSA), however, the invading 3’ strand is extended through polymerase activity but reanneals back to its original partner immediately after, creating a non-crossover product (as opposed to in NHEJ and DSBR). This is essential for preventing crossing over at G1/S and S/G2, for instance, as loss-of-heterozygosity (LOH) is a very frequent cause of carcinogenesis (Li et al, 2008).
In addition to its repair and cohesive functions, meiotic HR is a key driver of variation, as it gives rise to allele reshuffling, and as such it is an important driver of evolution.
Overall, there is still much discover about the exact mechanism of this process, but such knowledge can give us key insight into cancer therapy, as well as into tackling many other HR-related diseases.
Heyer, W.-D., Ehmsen, K.T. and Liu, J. (2010). Regulation of Homologous Recombination in Eukaryotes. Annual Review of Genetics, [online] 44(1), pp.113–139. Available at: https://www.deepdyve.com/lp/annual-reviews/regulation-of-homologous-recombination-in-eukaryotes-Yk02SbtCKj [Accessed 15 Oct. 2019].
Keeney, S. (2008). Spo11 and the Formation of DNA Double-Strand Breaks in Meiosis. Genome dynamics and stability, [online] 2, pp.81–123. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3172816/.
Li, X. and Heyer, W.-D. (2008). Homologous recombination in DNA repair and DNA damage tolerance. Cell Research, [online] 18(1), pp.99–113. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3087377/.
Liu, T. and Huang, J. (2016). DNA End Resection: Facts and Mechanisms. Genomics, Proteomics & Bioinformatics, [online] 14(3), pp.126–130. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4936662/ [Accessed 24 May 2019].
Son, M.Y. and Hasty, P. (2019). Homologous recombination defects and how they affect replication fork maintenance. AIMS Genetics, [online] 5(4), pp.192–211. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6690234/ [Accessed 8 Oct. 2020].
Symington, L.S. (2014). End Resection at Double-Strand Breaks: Mechanism and Regulation. Cold Spring Harbor Perspectives in Biology, [online] 6(8). Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4107989/ [Accessed 8 Oct. 2020].
Takeda, S., Hoa, N.N. and Sasanuma, H. (2016). The role of the Mre11–Rad50–Nbs1 complex in double-strand break repair—facts and myths. Journal of Radiation Research, [online] 57(Suppl 1), pp.i25–i32. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4990115/ [Accessed 8 Oct. 2020].
Moynahan, M.E. and Jasin, M. (2010). Mitotic homologous recombination maintains genomic stability and suppresses tumorigenesis. Nature Reviews Molecular Cell Biology, 11(3), pp.196–207.
Watson, J.D. and Al, E. (1987). Molecular biology of the gene / Vol. 1, General principles. Menlo Park, Ca ; London Etc.: Benjamin/Cummings, Cop.