Meiotic DSBs are not randomly distributed along eukaryotic chromosomes; rather, they are concentrated within discrete regions described as DSB hotspots [92]. To generate a high-resolution physical map of the meiotic DSB landscape, several methodologies have been established over the past three decades. SPO11-oligo mapping is achieved by immunoprecipitating tagged-Spo11 bound with oligonucleotides, which subsequently go through end-labeling, purification and sequencing [98]. This method has been effectively applied in several yeast species, mouse and Arabidopsis (Table 2) [92,99,100,101,102]. SSDS takes advantage of an antibody specifically recognizing either one of two DNA recombinases, RAD51 or DMC1, and utilizes chromatin immunoprecipitation to enrich single-stranded DNA that has undergone single-end invasion. With the high-throughput sequencing of enriched single-stranded DNA, the genome-wide distribution of DSB hotpots was successfully obtained in maize, mouse and human [103,104,105] (Table 2).

DSB hotspot designation is governed by a hierarchy of interrelated factors, including cis-regulatory elements, chromatin accessibility and high-order chromosome architecture (Table 2) [106]. Of particular prominence is the striking presence of DSB hotspots in repetitive DNA, such as transposon elements, although this aspect varies among species [10]. In maize, only ~25% of DSB hotspots have been discovered near genes, and the remaining hotspots are distributed in repetitive DNA—predominantly in Gypsy retrotransposons [105]. Although the Arabidopsis genome has fewer transposons compared with maize, a proportion of DSB hotspots were identified in Helitron/Pogo/Tc1/Mariner/MuDR DNA transposons [99]. Similarly, although most DSB hotspots tend to occur in genic regions, a significant overlap have been observed within the MULE-MuDR, TcMar-Mariner, hAT-Charlie and PiggyBac transposon families [107]. Moreover, DSB hotspots are evident in Ty retrotransposons in budding yeast. What is well established is that the meiotic recombination within or adjacent to repetitive elements is rare because this could lead to homologous recombination between non-allelic repeats, likely resulting in harmful chromosomal rearrangements and compromising genome stability [108,109]. Therefore, the paradox between the appearance of DSB hotspots and the suppression of recombination in these repetitive regions suggests that a mechanism exists to safeguard the repair of DSB in repetitive regions and prevent inadvertent crossing over.

CDKs in conjunction with their cyclin partners represent an ancient molecular switch that promotes and regulates cell-cycle progression [4,114,115]. The fundamental theme of how CDKs mediate meiotic recombination initiation was mostly drawn from studies in yeast. In S. cerevisiae, the activation of Cdc28 by its two B-type cyclin partners, Clb5 and Clb6, stimulates the phosphorylation on Mer2 during pre-meiotic DNA replication (Figure 1A) [115,116,117,118,119]; subsequently, phosphorylated Mer2 recruits other Spo11-accessary proteins to initiate DSB formation (Figure 1B) [45,120,121]. Similarly, in S. pombe, the association between Cdc2 and any of the three cyclins, Crs1, Cig1 and Cig2, is crucial for DSB formation [114].

In Arabidopsis, there are at least five types of cell-cycle CDKs (CDKA;1, CDKB;1, CDKB1;2, CDKB2;1 and CDKB2;2) and more than 50 cyclins [122,123,124,125,126,127,128,129,130,131], of which a few have been characterized as functioning during meiosis. CDKA;1 was originally identified as the key regulator for both mitosis and meiosis progression in Arabidopsis [129,131,132], and very recently, CDKA;1 was identified as a major regulator of meiotic recombination by mediating the number and placement of crossovers of homologous chromosomes [124,131]. However, CDKA;1 seems not to regulate meiotic prophase I although the exact influence on DSB formation was not determined in that study [124]. SDS, a cyclin protein, has been shown to control the formation of crossovers, but not DSBs, which differs from the necessity of its rice homologs in DSB formation [87,127,132,133]. TAM, an A-type cyclin (CYCA1;2), is indispensable for meiosis termination at the end of the first meiotic division [134,135]. Therefore, the core CDKs or related cyclins that directly regulate DSB formation remain to be explored in plants.

Once DSBs are made, there are several mechanisms to maintain the proper number of DSBs within a chromosome region [95,111,136,137,138,139,140,141]. This phenomenon, termed DSB homeostasis, prevents the severe effects caused by excess or insufficient Spo11 activity [136,137,140,142,143]. In S. cerevisiae, Tel1 and the related protein kinase Mec1 act as key factors that synergistically fine-tune the number of DSBs [141,142,144]. Lack of Tel1 leads to a considerable increase in the level of Spo11-oligonucleotide complexes, which are the by-product of meiotic DSB formation, indicating that Tel1 negatively regulates DSB formation (Figure 1C) [95]. In contrast, a defect in Mec1 activity causes a dramatic reduction in DSB number via inhibition of Ndt80, which triggers the exit from pachytene and shuts down Spo11-DSB catalysis of the cell cycle [145,146]. Mec1 ensures persistent Spo11 activity and indirectly promotes DSB formation, revealing that Mec1 positively regulates DSB formation (Figure 1C) [147,148]. In addition, recent studies in S. cerevisiae have revealed a cis-regulatory machinery, termed DSB interference, that reduces the frequency of coincident DSBs at the region adjacent to the preexisting DSB [136,142,143], and this process is dependent on TEL1 over short distances of ~70–100 kb (Figure 1D) [95,149]. On the contrary, the other machinery, known as trans-inhibition, defines the ability of a DSB to occur on a single chromosome, and this suppresses DSB designation on its homolog and sister chromatid at the same locus [141]. This mechanism relies on both Tel1 and Mec1, ensuring that an interhomolog interaction surrounding a DSB will not occur twice at the allelic or two nearby chromosomal positions so that DSBs are constrained to one per pair of homologs (Figure 1E) [141]. Interestingly, homolog engagement can also restrict the number of DSBs by inhibiting Spo11 activity (Figure 1E) [101,150]. Moreover, Tel1 and/or Mec1 can phosphorylate Rec114 directly, which limits its interaction with DSB hotspots and consequently reduces DSB formation genome wide (Figure 1C) [140].

In Arabidopsis, Tel1 and Mec1 homologs exist, termed ATM and ATR, respectively. However, ATM and ATR play synergetic roles in maintenance of genomic stability in meiotic cells by processing Spo11-dependent DSBs rather than influencing DSB formation [151,152,153,154,155]. Therefore, either ATM/ATR-mediated signaling pathway is evolutionarily divergent across different species or the effects of ATM/ATR on DSB formation is too subtle to be detected by conventional methods.