Homologous Recombination

Homologous recombination (HR) is a conserved enzymatic pathway that exchanges nucleotide sequences between two identical or nearly identical DNA duplexes. Across all domains of life, homologous recombination serves three primary biological functions: in proliferating somatic and microbial cells, it provides an error-free mechanism of DNA repair for DNA double-strand breaks (DSBs) and interstrand crosslinks (ICLs) as well as for restarting collapsed replication forks; in sexually reproducing eukaryotes, it establishes physical chiasmata between maternal and paternal chromosomes during Prophase I of meiosis while shuffling combinations of alleles; and in bacteria and archaea, it integrates exogenous linear DNA acquired through horizontal gene transfer. Because homologous recombination copies sequence information directly from an undamaged homologous template—preferentially the sister chromatid during the S and G2 phases of the cell cycle—it restores the original nucleotide sequence across a break without the insertions or deletions characteristic of non-homologous end joining.

Two-Tiered 5′→3′ End Resection

In eukaryotic cells, commitment to repair a double-strand break via homologous recombination rather than classical non-homologous end joining (c-NHEJ) is decided at 5′→3′ DNA end resection, which converts duplex ends into long 3′-terminated single-stranded DNA (ssDNA) tails that cannot be bound by Ku70/Ku80. Resection proceeds through a two-tiered enzymatic cascade:

1. Short-range resection (initiation): The MRN complex—comprising the nuclease MRE11, the ATPase RAD50, and NBS1 (MRX in Saccharomyces cerevisiae)—binds the DSB terminus and recruits ATM kinase. Activated by phosphorylated CtIP (Sae2 in yeast) and BRCA1–BARD1, MRE11 uses its endonuclease activity to nick the 5′-terminated strand 15 to 300 nucleotides internal to the DSB end and its 3′→5′ exonuclease activity to chew back toward the break, clearing protein-blocked termini and exposing a short 3′ overhang.
2. Long-range resection (extension): Two processive pathways extend resection over 1 to 5 kilobases in the 5′→3′ direction: the exonuclease EXO1, and the RecQ helicase BLM (Sgs1 in yeast) partnered with the endonuclease/helicase DNA2. The nascent 3′ ssDNA tail is immediately coated by Replication Protein A (RPA), which melts hairpins and protects the single strand.

Presynaptic Filament Assembly and Recombinase Mediators

Because RPA binds ssDNA tightly, it kinetically blocks the core eukaryotic recombinase RAD51 (homologous to bacterial RecA and archaeal RadA) from nucleating onto the DNA. Overcoming this barrier requires recombinase mediator proteins: in vertebrates, BRCA1–BARD1 recruits the bridging scaffold PALB2, which localizes BRCA2 (3,418 amino acids) to the ssDNA–dsDNA junction. BRCA2 contains eight conserved BRC repeats that bind monomeric RAD51 and a C-terminal OB-fold DNA-binding domain. Assisted by the five RAD51 paralogs (BCDX2 and CX3 complexes; Rad52 in yeast), BRCA2–PALB2 displaces RPA and nucleates RAD51 onto the 3′ ssDNA tail. In the presence of ATP, RAD51 polymerizes into a right-handed helical presynaptic nucleoprotein filament that stretches the underlying ssDNA to ~1.5 times the length of canonical B-DNA (~5.1 Å rise per nucleotide) in stacked nucleotide triplets.

Homology Search, Strand Invasion, and D-Loop Extension

The RAD51 presynaptic filament executes a genome-wide homology search and, assisted by the SWI/SNF-family ATPase translocase RAD54, invades an intact donor duplex (sister chromatid or homologous chromosome). During strand invasion, the invading 3′ ssDNA base-pairs with the complementary donor strand to form nascent heteroduplex DNA, while displacing the identical donor strand into a single-stranded displacement loop (D-loop). RAD54 then clears RAD51 from the invading 3′-OH terminus to prime repair synthesis catalyzed primarily by DNA polymerase δ (with PCNA and RFC).

Once the invading 3′ strand is extended by DNA polymerase δ inside the D-loop, two downstream pathways determine whether flanking markers undergo reciprocal crossover (CO) or remain non-crossover (NCO):

1. Synthesis-Dependent Strand Annealing (SDSA): In somatic mitotic cells, >90% of HR events follow SDSA. After D-loop synthesis extends past the break site, helicases (RTEL1, BLM, RECQL5, FANCM) unwind the D-loop and eject the extended 3′ strand, which anneals to the complementary 3′ overhang on the second DSB end for gap-filling synthesis and ligation. SDSA yields exclusively non-crossover (NCO) products, preventing somatic loss of heterozygosity (LOH).
2. Double Holliday Junction (dHJ) Pathway: If the displaced D-loop strand anneals to the second resected DSB end (second-end capture), synthesis and ligation covalently link the two duplexes via a double Holliday junction (dHJ):

• dHJ Dissolution (Non-Crossover): The BTR dissolvasome (BLM helicase – Topoisomerase IIIα – RMI1/RMI2) branch-migrates the two junctions together into a hemicatenane decatenated by TOP3A, producing 100% NCO products (loss of BLM causes Bloom syndrome).
• Endonucleolytic dHJ Resolution (Crossovers & Non-Crossovers): Structure-selective resolvases (GEN1 and the SMX tri-nuclease complex [SLX1–SLX4, MUS81–EME1, XPF–ERCC1]) cleave the four-way junctions to yield either NCOs or reciprocal crossovers (COs).

CDK-Mediated S/G2 Restriction vs. c-NHEJ in G1

Cells restrict homologous recombination to the S and G2 phases (when an identical sister chromatid is available) via cyclin-dependent kinases (CDK1/CDK2). In G1 phase, 53BP1 recruits RIF1 and the Shieldin complex (SHLD1–3/REV7) to DSB chromatin, blocking 5′→3′ resection and channeling breaks into Ku70/80-mediated c-NHEJ. In S and G2 phases, CDK2/CDK1 phosphorylate CtIP (Thr847) and EXO1 to activate resection, while BRCA1–BARD1 recognizes newly replicated H4K20me0 chromatin and displaces 53BP1–Shieldin.

Spo11-Programmed Breaks and Meiotic Crossovers

During Prophase I of meiosis, the transesterase Spo11 (Keeney, Giroux, and Kleckner, 1997) introduces 150 to 400 programmed DSBs per meiocyte (targeted in mammals by PRDM9). Resected ends load both RAD51 and the meiosis-specific recombinase DMC1, biasing invasion toward the homologous chromosome. While ~90% of meiotic DSBs resolve as non-crossovers, 1 to 3 DSBs per homolog pair are stabilized by the ZMM complex (MSH4–MSH5) and cleaved by MLH1–MLH3 (MutLγ) into obligate Class I crossovers (chiasmata).

Germline loss-of-function mutations in BRCA1, BRCA2, PALB2, RAD51C, and RAD51D cause Hereditary Breast and Ovarian Cancer (HBOC) and pancreatic/prostate cancers. As Hannah Farmer, Alan Ashworth, Helen Bryant, and Thomas Helleday showed in 2005, BRCA1/2-deficient tumors suffer synthetic lethality when treated with PARP1/2 inhibitors (olaparib, niraparib, talazoparib), which trap PARP1 on single-strand breaks that collapse into one-ended DSBs during S phase. In biotechnology, co-delivering programmable nucleases in CRISPR-Cas systems with a donor template stimulates local Homology-Directed Repair (HDR) by 1,000- to 10,000-fold for precision genetic engineering.

What is the difference between a crossover and gene conversion?

A crossover (CO) involves the reciprocal breakage and exchange of the flanking chromosome arms around a recombination site, swapping distal maternal and paternal segments. Gene conversion is the non-reciprocal copying of a short sequence tract (typically 50 to 1,000 bp) from an intact donor duplex into a broken recipient duplex during SDSA synthesis or heteroduplex mismatch repair, converting a heterozygous locus (A/a → A/A) without exchanging the flanking chromosome arms.

What is Break-Induced Replication (BIR)?

Break-Induced Replication (BIR) is a subpathway of homologous recombination triggered when only one end of a double-strand break shares homology with a donor template—such as at a collapsed replication fork or an eroded telomere in Alternative Lengthening of Telomeres (ALT) cancer cells. Driven by the DNA polymerase δ subunit POLD3 (Pol32) and Pif1 helicase, the invading 3′ strand establishes a migrating D-loop bubble that copies tens to hundreds of kilobases to the chromosome terminus.

How do BRCA1/2-mutant tumors acquire resistance to PARP inhibitors?

Under selective pressure from PARP inhibitors or platinum chemotherapy, BRCA1/2-mutant tumors evolve resistance through two primary mechanisms: (1) somatic reversion mutations—secondary intragenic deletions or indels in BRCA1 or BRCA2 that restore the open reading frame and functional protein; or (2) specifically in BRCA1-deficient tumors, somatic loss of 53BP1 (TP53BP1) or Shieldin subunits (REV7, SHLD1–3), which restores 5′→3′ DNA end resection and allows PALB2–BRCA2 to load RAD51 independently of BRCA1.

How does bacterial RecA differ from eukaryotic RAD51?

While bacterial RecA (~38 kDa) and eukaryotic RAD51 (~37 kDa) share a homologous central ATPase domain that polymerizes into a right-handed helical filament on single-stranded DNA, RecA hydrolyzes ATP ~100-fold faster and possesses an intrinsic co-protease activity: ssDNA-bound RecA* filaments stimulate autocatalytic cleavage of the LexA transcriptional repressor to induce the bacterial SOS DNA damage response (as well as cleaving phage λ cI repressor), a signaling role performed in eukaryotes by the ATR–Chk1 kinase cascade.

1. San Filippo, J., Sung, P., & Klein, H. (2008). Mechanism of eukaryotic homologous recombination. Annual Review of Biochemistry, 77, 229–257. doi:10.1146/annurev.biochem.77.061306.125255
2. Wright, W. D., Shah, S. S., & Heyer, W. D. (2018). Homologous recombination and the repair of DNA double-strand breaks. Journal of Biological Chemistry, 293(27), 10524–10535. doi:10.1074/jbc.TM118.000372
3. Farmer, H., McCabe, N., Lord, C. J., Tutt, A. N., Johnson, D. A., Richardson, T. B., ... & Ashworth, A. (2005). Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature, 434(7035), 917–921. doi:10.1038/nature03445
4. Keeney, S., Giroux, C. N., & Kleckner, N. (1997). Meiosis-specific DNA double-strand breaks are catalyzed by Spo11, a member of a widely conserved protein family. Cell, 88(3), 375–384. doi:10.1016/S0092-8674(00)81876-081876-0)