FIGL1 attenuates meiotic interhomolog repair and is counteracted by the RAD51 paralog XRCC2 and the chromosome axis protein ASY1 during meiosis

Introduction

During meiosis, the repair of DNA double-stranded breaks (DSBs) by homologous recombination (HR) yields crossovers (COs) and noncrossovers (NCOs) (Hunter, 2015; Wang & Copenhaver, 2018). Meiotic COs between homologous chromosomes (interhomolog) rather than between sister chromatids (intersister) serve important mechanical and evolutionary roles (Schwacha & Kleckner, 1994, 1997). The choice of the sister or nonsister chromatid template for repair is thus a key determinant for the outcome of meiotic recombination.

DNA strand exchange recombinases are central in regulating the choice of DNA template for DSB repair (Brown & Bishop, 2014; Humphryes & Hochwagen, 2014). RAD51 and DMC1 recombinases are two eukaryotic RecA homologs and can assemble into nucleofilaments on single-stranded DNA (ssDNA) generated from the processing of DSBs (Sheridan et al., 2008; Brown & Bishop, 2014). Both recombinases can perform homology searches of the genome and strand invasion on the donor template during meiosis. Cytologically, RAD51 and DMC1 form nuclear foci on meiotic chromosomes (Bishop, 1994; Kurzbauer et al., 2012; Brown et al., 2015; Slotman et al., 2020). Studies in many species contend that meiotic break repair occurs in two temporally distinct phases: a DMC1-permissive phase (Phase 1) followed by a RAD51-permissive phase (Phase 2) (Hayashi et al., 2007; Kim et al., 2010; Crismani et al., 2013; Enguita-Marruedo et al., 2019; Toraason et al., 2021; Ziesel et al., 2022). In the DMC1-permissive phase, DMC1 predominantly repairs DSBs and catalyzes interhomolog recombination, whereas RAD51 is kept catalytically inactive (Tsubouchi & Roeder, 2006; Busygina et al., 2008; Niu et al., 2009; Lao et al., 2013; Callender et al., 2016). In the RAD51-permissive phase, the RAD51-mediated pathway becomes active to repair remaining DSBs, mainly using sister chromatids (Crismani et al., 2013; Enguita-Marruedo et al., 2019; Toraason et al., 2021; Ziesel et al., 2022). The RAD51-dependent pathway also repairs DSB on sisters before meiotic entry (Joshi et al., 2015).

During Phase 1, different RAD51-inhibiting strategies appear to have evolved in eukaryotic species. The Mek1-mediated pathway downregulates Rad51-dependent repair in budding yeast (Niu et al., 2009; Callender et al., 2016). This regulation is, however, not conserved in plants, because RAD51 can repair breaks in the absence of DMC1, albeit using sister chromatids inferred from a lack of interhomolog COs (Couteau et al., 1999; Wang et al., 2016). The mere presence of DMC1 attenuates RAD51 repair in yeast and Arabidopsis (Lao et al., 2013; Da Ines et al., 2022). Constitutive activation of Rad51, in addition to active Dmc1, elicits a longer repair time in yeast (Ziesel et al., 2022). Wild-type (WT) DMC1-mediated interhomolog recombination nonetheless requires the presence of RAD51, but not its catalytic activity (Cloud et al., 2012; Da Ines et al., 2013b). In Arabidopsis, RAD51 fused to GFP at its C-terminus (RAD51-GFP) is catalytically inactive and unable to repair breaks in mitotic and meiotic cells but supports DMC1-mediated repair during meiosis (Da Ines et al., 2013b). This function suggests that the catalytic activity of RAD51 is nonessential for DMC1-mediated repair in plants.

In vivo functions of DMC1 and RAD51 require many accessory proteins in eukaryotes. In plants, BRCA2 mediates the formation of RAD51 and DMC1 foci, whereas SDS is specifically required for DMC1 focus formation (Azumi, 2002; Seeliger et al., 2012; Fu et al., 2020). These mediators appear to act in vivo at the nucleofilament formation step. Further, MND1 and HOP2 are two evolutionarily conserved proteins required for the DNA exchange activity of DMC1 in plants (Petukhova et al., 2005; Chan et al., 2014; Uanschou et al., 2014). In mnd1 and hop2, DMC1 hyperaccumulation inhibits meiotic DSB repair in Arabidopsis (Kerzendorfer et al., 2006; Panoli et al., 2006; Vignard et al., 2007; Stronghill et al., 2010; Farahani-Tafreshi et al., 2022). However, weak DMC1 activity in the Arabidopsis hop2-2 hypomorphic mutant greatly compromises interhomolog repair and allows RAD51-dependent DSB repair on sisters (Uanschou et al., 2014). RAD54 is also required for RAD51 functions but is unnecessary for meiotic DSB repair in the presence of DMC1 in Arabidopsis (Hernandez Sanchez-Rebato et al., 2021). Further, Arabidopsis has five structurally related RAD51 paralogs: RAD51C, XRCC3, RAD51D, RAD51B, and XRCC2 (Bleuyard et al., 2005). These paralogs can form a tetrameric complex called the BCDX2 complex (Osakabe et al., 2002). Arabidopsis RAD51C and XRCC3 are essential for RAD51 focus formation and meiotic repair (Bleuyard & White, 2004; Abe et al., 2005; Vignard et al., 2007), but RAD51B, RAD51D, and XRCC2 play no critical role in RAD51-dependent meiotic DSB repair, irrespective of the presence or absence of DMC1 (Bleuyard et al., 2005; Hernandez Sanchez-Rebato et al., 2021). However, the loss of Arabidopsis RAD51B and XRCC2 slightly increases the meiotic recombination rate, implying their as yet unascertained roles in meiotic repair (Da Ines et al., 2013a).

Meiotic chromosome axis proteins ensure DSB repair and CO formation between homologs in a process called interhomolog bias (Morgan et al., 2023). Arabidopsis ASY1, ASY3, and ASY4 are three meiotic axis-associated proteins that promote synapsis, a process allowing tethering between homologs through polymerization of synaptonemal complex (SC) proteins such as ZYP1 (Caryl et al., 2000; Armstrong et al., 2002; Higgins et al., 2005; Sanchez-Moran et al., 2007; Ferdous et al., 2012; Chambon et al., 2018; Vrielynck et al., 2023). ASY1 localizes on the meiotic axis in an ASY3-dependent manner and is depleted from synapsed regions, following SC assembly between homologs (Ferdous et al., 2012). Loss of ASY1, ASY3, and ASY4 results in a substantial reduction in interhomolog COs, albeit at different magnitudes, with meiotic DSBs predominantly repaired on sisters. ASY1 is required for DMC1 stabilization, suggesting a functional relationship between the meiotic axis and repair machinery (Sanchez-Moran et al., 2007). How meiotic chromosome axis proteins promote DSB repair between homologs is currently unclear.

Most eukaryotes have two classes of COs formed between homologs. In Arabidopsis, class I constitutes a significant proportion (85–90%) of COs, mediated by the ZMM group of proteins (SHOC1, PTD, HEI10, ZIP4, MSH4/5, and MER3) and MLH1/3 (Mercier et al., 2015). The class I CO pathway ensures the obligate CO between homologs but is sensitive to CO inference that avoids the formation of additional class I COs in close proximity (Wang et al., 2015; Lloyd, 2023). Class II COs are derived from the structure-specific endonuclease-dependent pathway, including MUS81 (Berchowitz et al., 2007; Wang & Copenhaver, 2018). Three nonredundant anti-class II CO pathways involving FANCM, RECQ4A & RECQ4B, and FIDGETIN-LIKE-1 (FIGL1) limit meiotic COs through distinct mechanisms (Crismani et al., 2012; Girard et al., 2014, 2015; Séguéla-Arnaud et al., 2015, 2017; Fernandes et al., 2018a). Although these three pathways regulate class II COs, FIGL1 can also control the distribution of class I COs among chromosomes (Fernandes et al., 2018a; Li et al., 2021). Arabidopsis mutants lacking FIGL1 show a moderate increase in COs with occasional achiasmatic chromosomes (Girard et al., 2015; Fernandes et al., 2018a).

FIGL1 is a member of the AAA-ATPase family and has enigmatic roles in positively and negatively regulating meiotic CO formation. Arabidopsis and wheat FIGL1 limit class II COs potentially by preventing aberrant recombination intermediates or chromosome associations (Girard et al., 2015; Fernandes et al., 2018a; Kumar et al., 2019; Osman et al., 2024). FIGL1 and its mammalian ortholog FIGNL1 physically interact with the two recombinases and can antagonize positive mediators of RAD51/DMC1, such as BRCA2 and SDS in Arabidopsis or SWSAP1 in humans (Girard et al., 2015; Fernandes et al., 2018a; Kumar et al., 2019; Matsuzaki et al., 2019). Surprisingly, maize FIGL1 can act cooperatively with BRCA2 to positively regulate meiotic recombination (T. Zhang et al., 2023). Arabidopsis figl1, rice figl1, and mice fignl1^cko mutants show a change in the dynamics of RAD51 and DMC1 foci, leading to the deregulation of strand invasion, which supports a conserved role of FIGL1/FIGNL1 in meiotic DSB repair (Zhang et al., 2017; Fernandes et al., 2018a; Yang et al., 2022; Ito et al., 2023; Q. Zhang et al., 2023). FIGNL1 is also involved in DSB repair via homologous recombination in somatic cells (Yuan & Chen, 2013; Matsuzaki et al., 2019). FIGL1/FIGNL1 is thus a conserved regulator of RAD51- and DMC1-mediated strand invasion and may function in the fine-tuned regulation of the strand invasion step to promote accurate meiotic DSB repair. How FIGL1 regulates the outcome of meiotic break repair when RAD51- and/or DMC1-dependent pathways are fully or partially impaired remains unknown. In this study, we investigated the impact of the functional relationship of FIGL1 with components of HR repair machinery and chromosome axis genes on the outcome of meiotic DSB repair. We demonstrate that these genetic interactions are an essential determinant of meiotic break repair outcomes.