The dynamic association of SPO11-1 with conformational changes of meiotic axial elements in maize

Double strand breaks (DSBs) are generated at the beginning of meiosis by the evolutionarily conserved SPO11 complex that targets preferentially open chromatin. However, how DSBs are formed with respect to chromosome axes in a timely controlled manner remain unclear. Here, we analyzed the maize spo11-1 mutant and show that it strongly impairs DSB and bivalent formations. Notably, cytological characterization in the spo11-1 mutant revealed abnormally twisted axial elements that persisted until pachytene. In contrast, examinations of precisely staged wild-type meiocytes uncovered a transient remodeling of axial elements, changing from a curly to linear morphology during leptotene to zygotene transition, which is coincident with DSB formation. Using a SPO11-1 antibody, approximately 300 foci were detected from leptotene to pachytene in wild-type meiocytes. Interestingly, when examining distances between SPO11-1 foci to chromosome axes by super-resolution microscopy, predominant loading of SPO11-1 onto axial elements occurs concordantly with alteration of axial elements during leptotene. Taken together, our results suggest a dynamic localization of SPO11-1 during early meiosis that is correlated with a remodeling of the axial element conformation. Author Summary Meiosis is an important step in sexual reproduction for producing haploid gametes. Recombination is the most crucial step during meiotic prophase I, which enables pairing of homologous chromosomes prior to their reductional division, and generates new combinations of genetic alleles for transmission to the next generation. Meiotic recombination is initiated by generation of many DNA double-stranded breaks (DSBs), and their repair leads to reciprocal exchanges between paternal and maternal chromosomes. The activity, timing and location of this DSB machinery must be controlled precisely, but a full picture is still obscure. By studying maize SPO11-1, a topoisomerase-related enzyme required for making DSBs, we uncovered dynamic localization of SPO11-1 that is correlated to a transient structural change of the axial elements of wild-type plant meiocytes by super-resolution microscopy. Interestingly, loss of SPO11-1 impairs this structural change, resulting in a prolonged spiral morphology. Our study suggests an intimate relationship between SPO11-1 and axial elements and has implications for our understanding of how DSB formation is controlled in plants.


Introduction
Meiosis is a special type of cell division that includes two successive nuclear divisions after one round of DNA replication and it allows segregation of only one copy of homologous chromosomes to halve the ploidy number. A faithful segregation of chromosomes requires meiotic recombination during prophase I, which mediates pairing of homologous chromosomes and creates physical connections until their segregation (1). Meiotic recombination is initiated by the introduction of double strand breaks (DSBs) into chromosomal DNA by SPO11, a protein belonging to a distant topoisomerase VI family (2). Absence of SPO11 results in defective initiation of meiotic recombination in all species analyzed, leading to sterile phenotype.
The self-inflicted DNA breaks introduced by SPO11 are tightly controlled both temporally and spatially to allow proper initiation of recombination (1). These DSBs must be repaired to ensure genome integrity. Only a small subset of DSBs produces crossovers (COs) that involve long reciprocal exchanges between homologous chromosomes, whereas the residual DSBs lead to non-crossovers (NCOs) (3). These CO sites, seen as chiasmata at diplotene stage, ensure that the pairs of homologs, i.e. the bivalents, maintain their association and align properly at the metaphase small chromosomes (35)(36)(37)(38)(39)(40). Here, we show that maize Spo11-1 is, as expected, required for DSB and bivalent formation. Interestingly, cytological analysis of the spo11-1 mutant exhibited an unexpected phenotype, with AEs exhibiting a spiral morphology from leptotene through late zygotene. In contrast, in wild-type meiocytes, a transient AE conformational change was observed during late leptotene, which coincides with an increase of SPO11-1 number associated with AEs. These findings provide the first cytological evidence of dynamic SPO11-1 localization during DSB formation and suggest that a possible SPO11-1 dependent mechanism may orchestrate changes in AE structures.

Results
The two maize meiotic mutants mtm99-14 and mtm00-03 are allelic We used a forward genetic approach to identify genes involved in the control of the maize meiotic program based on sterility and meiotic cytological analysis (25,41,42). Among the meiotic mutants, the mtm99-14 and mtm00-03 mutants exhibited severe male and female infertility (Fig 1A and 1B and mtm00-03 mutants were manually pollinated with WT pollens, both mutants produced one or two kernels (0.17 % of 1152 ovules for mtm99-14 and 0.07% of 1421 ovules for mtm00-03). We found that mtm99-14 and mtm00-03 are allelic based on segregation for the male and female sterile phenotype among the F1 progeny in an allelism test. The heteroallelic mtm99-14/mtm00-03 plants are male sterile and manual pollination of their female ears with WT pollens produced no or few kernels (0.12% of 1624 ovules).
We then analyzed meiotic progression of homozygous mtm99-14, mtm00-03 and heteroallelic mtm99-14/mtm00-03 mutants in detail. In the WT, ten bivalents align on the metaphase plate during metaphase I, but mutants contain mainly univalents with 5-7% of meiocytes forming one or few bivalents (Fig 1C,upper panel,and 1D). In contrast to normal tetrad cells in WT, abnormal microspores are seen as unequal tetrads, as tetrads with micronuclei, or as polyads in mutants (Fig 1C,lower panel,and 1E). This phenotype indicates that both mtm99-14 and mtm00-03 exhibit impaired bivalent formation and the resulting univalents randomly segregate, leading to the formation of unequal spores and inviable pollens.
The strong sterility of maize ears indicates that female meiosis is affected similarly to that of males. That few seeds were obtained from homozygous mutants when pollinated with WT pollens indicates that these mutants can only produce rare viable female megaspores. In maize asynaptic mutants, the probability of obtaining viable spores with the full haploid set of ten chromosomes from random segregation is 0.12% (1/2 10 ). The proportion of seeds obtained from these mutants pollinated by WT pollens is not statistically different from the expected proportion of completely random segregation (ε= 0.82 for mtm99-14, ε = 0.32 for mtm00-03, ε = 0.33 for mtm99-14/mtm00-03 < ε α 5% =1.96). This analysis demonstrates that the male and female infertility of mtm99-14 and mtm00-03 is due to severe defects of bivalent formation.

mtm99-14 and mtm00-03 mutants exhibit deletions of the Spo11-1 locus associated with Mu insertions
The mtm00-03 mutant was first mapped to chromosome 5 by molecular markers (see Material and Methods). To identify the gene, we used the Mu-Illumina strategy to analyze Mu insertion borders that cosegregate with sterile phenotype (43) and found that mtm99-14 mutant contains Mu element sequences inserted into a Tourist transposable element that is located 9 kb upstream of the Spo11-1 gene (GRMZM2G129913/ZM05G04150) on maize chromosome 5.
Similarly, mtm00-03 mutant carries a Mu insertion about 4 kb upstream of the Spo11-1 gene (Fig  2A). These Mu sequences in both mutants are associated with deletions, which likely resulted from Mu-induced excisions. Both deletions span the Spo11-1 and phytochrome C2 genes (PhyC2: GRMZM2G129889/ZM05G04160). We could not determine the lengths of deletions and only one Mu border was recovered in our Mu-Illumina sequence analyses. Nevertheless, reverse transcription (RT) PCR experiments confirmed that neither of the mtm99-14 or mtm00-03 mutants expressed the Spo11-1 gene in meiotic anthers (Fig 2B). Upon examining PhyC2 expression by RT-PCR, we detected an unexpected signal in both mutants ( Fig S1). Further sequence analysis showed that the resulting DNA fragment in the mtm99-14 and mtm00-03 mutants was amplified from the paralogous PhyC1 (GRMZM2G057935/ZM01G52970), which is located on chromosome 1 and shares 94% identity with the PhyC2 coding region. Given this high similarity, our primers were unable to distinguish the PhyC1 and PhyC2 coding sequences.
Considering the complex rearrangements that may occur in both mtm99-14 and mtm00-03, we searched for other mutations affecting only the Spo11-1 gene. From the Trait Utility System for Corn (TUSC) transposon mutagenesis population (44), one additional mutant, named spo11-1-1 contains a Mu insertion located in the first exon (Fig 2A) and exhibits male and female infertility. Cytological analyses of male meiosis in this mutant also exhibited 20 univalents in most meiocytes, and similarly to mtm99-14 and mtm00-03, 5% of meiocytes exhibited one or two bivalents (Fig 1D and S2). Allelism tests between mtm99-14/+ or mtm00-03/+ and spo11-  Fig 2B) and found that the variability arose from retention of the second intron ( Fig 2D). Both transcripts are predicted to produce functional SPO11-1 proteins because they contain a nuclear localization signal (KLRR) in the N terminal region, the topoisomerase-primase 'TOPRIM' domain that is essential for catalytic activity, and the conserved tyrosine residue in the winged-helix DNA-binding (WHD) domain ( Fig 2E). The longer transcript (named Spo11-1β) encodes a 426 amino acid protein that contains an extra 43 amino-acid domain translated from the intron 2 ( Fig 2D, 2E, S3). This additional region is predicted to form an alpha-helix motif upstream of the TopVIA type endonuclease domain ( Fig S4).
We designed transcript-specific primers (shown in Fig 2D) to examine temporal expression patterns of different transcripts. Quantitative PCR results revealed that the Spo11-1β transcript is about twice as abundant as the Spo11-1α transcript in leptotene/zygotene anthers, whereas Spo11-1α is expressed slightly more before prophase I ( Fig 2C). In order to understand whether both splicing variants can produce SPO11-1 protein isoforms, we generated a SPO11-1 antibody using an N-terminal peptide (aa 9-36). Using proteins extracted from meiotic anthers of WT and spo11-1-Δ03 plants, we detected two weak bands (arrowheads in Fig 2F) between 35-40 kDa in WT, which are absent in the mutant samples ( Fig 2F. The entire blot is shown in the Fig   S5). This result suggested that two SPO11-1 isoforms are produced in meiotic anthers.
Unfortunately, our attempts to further examine these two isoforms by generating isoform-specific antibodies were not successful.

The spo11-1 maize mutants are defective in DSB formation
Since SPO11-1 is required to initiate homologous recombination by creating DSBs, we determined cytologically whether the spo11-1 mutants are defective for meiotic DSBs. The terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was used to detect DNA free ends associated with DSBs. Compared to obvious signals detected in WT meiocytes at zygotene, we found that most of spo11-1 meiocytes did not show any TUNEL signals on chromatin ( Fig 3A). Only around 10% of the mutant meiocytes had a few TUNEL foci usually located in one region of the nucleus (n=7/74: 9.5% for spo11-1-Δ14, n=2/24: 8.3% for spo11-1-Δ03 and n=5/42: 11.9% for spo11-1-1) ( Fig S6). These few DSBs may explain the rare observation of bivalents in maize spo11-1 mutants (Fig 1D and S2), suggesting that a minor SPO11-1 independent pathway can lead to rare bivalent formations in maize.
To further confirm DSB formation defects, we applied antibodies against γH2AX, Replication Protein A1 (RPA1a), and RAD51 proteins. In various eukaryotes, phosphorylation of the histone variant H2AX (γH2AX) is induced rapidly at DSB sites. Then, RPA1 interacts with the single-stranded DNAs after DSB resection and, later, loading of the recombinase RAD51 facilitates recombination. In WT male meiocytes, γH2AX signals appeared at early zygotene, observed as numerous dots on maize chromosomes. In contrast, most of spo11-1 meiocytes exhibited no γH2AX signals ( Fig 3B). Numbers of RPA1a and RAD51 foci in WT peaked during mid-zygotene, whereas most of the mutant meiocytes completely lacked signals (Fig 3C and 3D).
Maize spo11-1 mutants present an aberrant, twisted axial element structure In maize, meiosis in anthers of similar sizes is almost synchronous during prophase I, and the relative positions of anthers on a maize tassel are correlated to their developmental stages.
Thus, anther sizes and their positions on a tassel provide a relative time course for meiotic progression, i.e. meiosis in 2 mm anthers is more progressed than it is in 1.5 mm anthers on the same tassel. To follow recombination along meiotic progression, we dissected anthers from the same tassel in order and used anthers of the same size to prepare slides for immunostaining.
Therefore, our analyses provide adequate temporal resolution, which allows a detailed characterization of the course of recombination initiation.
Previous studies have shown that maize meiotic chromosome axes labeled with DSY2 and ASY1 antibodies appear as linear central cores along chromosomes at zygotene (29).
To investigate this phenotype in detail, we examined axes morphology at very early stage when axial elements first appear and enclose a nucleolus in the center. We noticed that at early leptotene, AEs exhibited a similar curly appearance in both WT and spo11-1 mutants ( Fig 4A).
The twisted axes are more evident when visualized through z-sections (Movies S1, S2). At late leptotene/early zygotene, when nucleolus is offset inside the nucleus, AEs in WT meiocytes become smoother and linear in most parts of chromosomes. However, unlike WT AEs that switch to a linear structure at late leptotene, twisted AEs persist in spo11-1 mutants (Figs 4B, 4C).

Synaptonemal complex is assembled promiscuously in spo11-1 mutants
In WT maize, upon telomere clustering, synapsis takes place mostly from telomere clustering regions and extends toward to interstitial regions as observed by the transverse filament protein ZYP1 signals (Fig 5A). At pachytene, ZYP1 is loaded along the entire length of synapsed chromosomes (Fig 5C). In contrast, ZYP1 is discontinuously installed first as short stretches in spo11-1 meiocytes ( Fig 5B). Until late pachytene, longer but discontinuous ZYP1 signals were observed in spo11-1 mutants ( Fig 5D).
Using transmission electron microscopy of silver nitrate stained SCs, we observed a similar conformation of abnormal axes in unsynapsed regions in the spo11-1 mutant (Fig 5F,   5G). In addition, some regions showed obvious promiscuous synapsis. For example, SC is formed within the same axis (arrows in Fig 5f', 5g'). This finding is consistent with short and discontinuous stretches of ZYP1 signals in the immunostaining result ( Fig 5D).

Dynamic SPO11-1 localization occurs concordantly with conformational changes of axial elements
To examine SPO11-1 localization, WT meiocytes from early leptotene to pachytene were subjected to immunolocalization. At early leptotene, most of SPO11-1 signals are located around nuclear periphery (Fig 6A, movie S3). Notably, chromatin stained by DAPI at this stage appears curly morphology, which is similar to axial element stained by DSY2 antibody ( The aberrant AE structure in the absence of SPO11-1 suggested an intimate relationship between SPO11-1 and AEs. To understand the relative dynamics of SPO11-1 and chromosome axes, SPO11-1 and DSY2 localizations were examined by structured illumination microscopy. On average, about 250-350 foci of SPO11-1 were detected during early prophase I in WT (blue boxes in the Fig 6H). In contrast, less than 50 foci were detected in spo11-1-Δ03 mutant.
Although the total number of SPO11-1 foci remained similar during WT leptotene to pachytene, their locations relative to chromosome axes seemed to be dynamic when imaged by super . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
To investigate their distributions further, we used the software ImageJ to calculate a "distance map" of DSY2 signals in Z stacks and then determined the distances between each SPO11-1 focus to the closest DSY2-labeled axes in 3D space (see material and methods). Among a cohort of meiocytes from an anther during leptotene, we observed leptotene meiocytes with no sign of homologous pairing (e.g. Fig 6E), as well as late leptotene meiocytes with obvious telomere bouquet formation (e.g. Fig 6F). Perhaps, this transition occurs rapidly. Nevertheless, we analyzed leptotene, zygotene and pachytene meiocytes and calculated the numbers of SPO11-1 foci associated with AEs (red boxes in Fig 6H). The leptotene meiocytes showed the highest number of AE-associated SPO11-1 foci, numbers of which gradually decreased in the zygotene and pachytene stages. Furthermore, the highest number of AE-associated SPO11 foci was detected when the AEs become smooth and linear (S2 Table). These results suggest a dynamic localization of SPO11-1 during recombination initiation, which coincides with a conformational change of axial elements.

Discussion
Although maize spo11-1 mutants exhibited the same meiotic sterility as observed in other species, thanks to the remarkable advantages of maize for cytological observations, our analyses have led to several unexpected results that provide new insights into DSB formation.
The most interesting finding of our study is the twisted AEs observed in the maize spo11-1 mutant, which prompted us to further examine AE morphology relative to DSB formation. Using carefully staged WT anthers, our cytological analysis allowed us to visualize a change in AE conformation during the leptotene to early zygotene. Initially at early leptotene, chromatins are organized into curly fibers by axis proteins that appear curly and surround a nucleolus in the center.
At this stage, SPO11-1 is mostly located on the nuclear periphery. The twisted AEs observed in maize may reflect the chromosome 'spiralization' phenomenon that was described in classical cytological studies during the 'spiral stage' of leptotene entry (45). However, this spiral stage is only obvious in species such as lilies or grasshoppers that have huge chromosomes and an extended meiotic prophase (46-48). Even in maize, it has not been reported previously. We interpret the fact that these spiral axes are more easily observed in the mutants than in WT because spo11-1 mutants have a prolonged leptotene-like stage with curly axes.
It has been reported that the leptotene-zygotene transition (prezygotene) in maize undergoes a dramatic change, including an increase in total chromosome volume, and a peripheral localization and clustering of telomeres (35). Meiotic DSBs are thought to form at this transitional stage (49). Here, we found that when meiocytes exhibited an offset nucleolus, numerous SPO11-1 foci are located inside the nucleus. Then, AEs gradually became smoother and more linear. The remodeling of AE linearization occurs prior to SC assembly in WT, since ASY1 signals that label unsynapsed regions are linear and smooth during WT zygotene ( Fig 4C). The presence of curly AEs is not dependent on SPO11-1, since spo11-1 mutants showed similar curly AEs at the beginning of leptotene. Although the biological role of this transient change in AE morphology (from curly to linear axes) is not yet clear, given the fact that the curly and twisted appearance of AEs persists in maize spo11-1 mutants, we postulate that the timely remodeling of AEs into linear structures depends on SPO11-1.
As expected, DSB formation is severely affected in maize spo11-1 mutants, leading to male and female sterility. The fact that SPO11-1 is required for DSB formation raises the possibility that remodeling of AEs into linear structures may result from DSB occurrence. However, abnormally twisted AEs have not been reported for other maize mutants that are defective in DSB formation, such as the dsyCS mutant (50). Another possibility is that loading of SPO11-1 onto AEs is related to the remodeling of AEs into linear structures. This scenario somewhat supported by our observations via super-resolution microscopy of dynamic SPO11-1 localization. Interestingly, our analyses showed that leptotene meiocytes have the highest proportion of SPO11-1 foci associated with AEs ( Fig 6F). Furthermore, the highest number of AE-associated SPO11-1 signals was found at late leptotene when the AEs become smooth and linear (S2 Table). Nevertheless, mutants in which only DSB formation is disrupted, but SPO11-1 localization is not affected, may be valuable for understanding the nature of AE remodeling.
Another interesting and interlinked finding here was the dynamic nature of SPO11-1 localization. Maize leptotene chromatin is about 400 nm wide and AEs are about 50 nm wide, running along the leptotene chromosomes. Together with our protocol that allows better preservation of the 3D organization of meiotic nuclei, and uses of structured illumination super-resolution microscopy, we were able to resolve SPO11-1 localization in detail. We showed that SPO11-1 appears first on nuclear envelope and then enters nucleus during leptotene. It is first located mostly on chromatin and then loaded on AEs at late leptotene. After zygotene, most SPO11-1 signals are present in nuclei, but not on AEs. To understand whether loading of SPO11-1 on AEs coincides with DSB formation, we sought to examine γH2AX and SPO11-1 localization.
Unfortunately, these antibodies were made in rabbit and proved unsuitable for this purpose. Future investigation using dual immunolocalization will provide a better understanding of SPO11-1 protein localization and DSB formation.
Our observations using super resolution microscopy have revealed for the first time in any organism the dynamic localization of SPO11-1 relative to AEs. We interpret this phenomenon with regard to the early recombinosome that tethered to axes, i.e. the "tethered loop-axis" model proposed in yeast (5, 15). Perhaps, the curly AEs play an active role in this process. For example, it permits or promotes the tethering process. Later, recombinosomes associated with AEs may be required to activate the DSB machinery. This scenario may explain the defective DSB formation observed in many AE mutants in various species.
Remodeling of AEs into linear axes seems to take place after SPO11-1 loading onto axes or DSB formation in WT. Perhaps, the remodeling of AEs into linear structure prevents more tethering events or additional DSB formation or is related to a reinforcement of inter-homolog recombination bias. Interestingly, yeast Pch2 has been shown to have Hop1 remodeling activity in vitro (51). Moreover, yeast Pch2 was suggested to coordinate with the checkpoint protein Tel1 ATM to control early DSB formation and to mediate homolog bias (52). In Arabidopsis, PCH2 localizes to AEs and is required for normal ASY1 depletion from chromosome axes during SC assembly (31). Although the relationship between DSB formation and AE functions is not yet clear, our study has demonstrated that in the absence of SPO11-1, AE morphology is distinct from that in WT, suggesting an intimate relationship between SPO11-1 and AEs. Our results not only provide the first cytological evidence to support the 'loop-tethered axis' model (5), but also shed lights on how DSB formation is controlled in plants.

Maize material
Maize inbred lines B73, A344, A632 and Mo17, as well as wild type siblings of the Mu insertional mutants and spo11-1 mutant alleles were grown in the greenhouse of Academia Sinica and previously in the UC Berkeley greenhouse and experimental field. The mtm99-14 / . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

Genetic mapping and Mu insertion determination.
To map the mtm00-03 mutation, bulk segregation analysis (BSA) of two mtm00-03 F2 population created by crossing the mtm00-03 line with inbred A344 or B73 were used as previously described (53). Sterile plants show a tight linkage to polymorphic markers IDP8472 and particularly IDP4279 that map to the end of the short arm of chromosome 5 on Bin 5.01 between physical positions 5.3 and 8.5 Mb. For each mutant, Mu insertions border sequences that cosegregate with the phenotype were determined using the Mu-Illumina border determination protocol as previously described (43).

RT-PCR analysis
RT-PCR analyses were conducted on RNA extracted using Trizol from meiotic anthers, as described (54). The Afd1 252bp cDNA, the PhyC 444bp cDNA fragments, and 3' and 5' portion of Spo11-1 transcripts were amplified using specific primers (S1 Table). The Phyc1 and Phyc2 presented as two distinct bands from genomic DNA but with an identical size to cDNA. All PCRs were performed using DMSO and the Invitrogen GoTaq kit with 30 cycles using Tm= 52°C but for the SPO11-1 5' couple where Tm=60°C. Quantitative PCR of Spo11-1α and Spo11-1β transcripts was performed using FastStart Universal SYBR Green Master mix with specific primers (S1 Table). Cyanase was used as an internal normalization control. PCRs were performed on an ABI 7500 real-time PCR system in a 64-well reaction plate. The RT-qPCR . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018; reaction was performed in triplicate for each RNA sample. Specificity of the amplification products and Ct values were analyzed using ABI prism dissociation curve analysis software.

Antibody production
To generate anti-SPO11-1 antibody, we used a peptide corresponding to amino acids 9-36 of SPO11-1 protein (RAAPLEGDEQQLRRRLEEAALLLRRIKG) as antigen to generate a rabbit polyclonal antibody, followed by affinity purification (LTK BioLaboratories, Taiwan).

Western blot
Anthers were dissected from immature tassels after their meiotic stages and sizes were determined and snap frozen in liquid nitrogen. Protein extraction was performed as previously described (29). Two hundred micrograms of proteins were loaded in a 20 cm long, 10% SDS-PAGE gel, run for 18 h at 100 V, and then transferred to a polyvinylidene fluoride (PVDF) membrane. The PVDF membrane was blocked overnight with 5% milk in Tris-buffered saline-Tween (TBST) solution (containing 1X TBS and 0.1% Tween 20) at 4°C. Subsequently, the PVDF membrane was incubated with anti-SPO11-1 antibody (1:1000 dilution in 5% milk/TBST) and then with the horseradish peroxidase conjugated goat anti-rabbit IgG secondary antibody (1:5000 dilution in TBST). After a TBST wash, signals were detected using the UVP Biospectrum 600 imaging system.

Immunolocalization
Immature tassels were carefully removed from 4-to 6-week-old plants and kept in moist paper towel until dissection or fixation. Anthers were dissected after their meiotic stages and sizes were determined and their positions on a tassel were recorded as a relative time course. We . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018; analyzed meiosis in synchronized cohorts of meiocytes in anthers using characteristics described previously (35) to precisely stage the progression of maize meiotic prophase in WT and mutant cells. Immunostaining was performed as described (55)

Image processing
The super-resolution z-stacking images were viewed with ZEN2012 software (Zeiss) to generate 2D projected images. To measure the distance between SPO11-1 and chromosome axes, the tool "ImageJ-FIJI" was used to generate SPO11-1 and DSY2 segmented 3D stack files. The 3D distance map was created using the plugin "3S Image Suite" (56), and distances between each SPO11-1 signal and nearby DSY2-labeled axes were analyzed using "3D manager" tool in 3D Image Suite.

TUNEL assay and TEM
The TUNEL assay for DSB detection was performed as described (57). First, meiocytes were embedded in polyacrylamide pads as described in the immunostaining protocol above. The TUNEL reaction was performed in a humid chamber for 2.5 h at 37°C with the in Situ Cell Death Detection Kit (Roche) using 100 μL of reaction mixture per slide. Slides were mounted and imaged with a Deltavision core microscope. TEM analysis was performed as described (57, 58).
. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

Authors Contributions:
AR wrote the paper, designed and carried out field, molecular experiments and immuno-cytological analyses. IG did the original forward genetic screen, determined the allelism between mtm00-03 and mtm99-14 and conducted initial cytological analyses of the spo11-1 mutants. JCK generated and analyzed SPO11-1 antibody, performed immunolocalization and western blot, and carried out field works in Taiwan. DHL performed super-resolution microscopy and analyzed curly axial element in spo11-1 mutant. LT made crosses, determined the allelism between mtm00-03 and mtm99-14, conducted the TEM analysis and provided the DNA of the mtm00-03 mutants for Mu-Illumina sequencing. AKGA did the RT-PCR analysis. YHK performed quantitative PCR for alternative spicing. KK did the BSA mapping and the RT-PCR analysis. RWC and AB did the Mu-Illumina border sequencing for mtm99-14 and mtm00-03. RM provided the original seed stocks of the TUSC lines (spo11-1-1). ZC initiated the project. CJRW co-wrote the paper with AR and supervised the project.

Acknowledgments
We thank Jihyun Moon for advises on BSA mapping analysis. We thank Sidae Lee and Angel Jung for help with specimen collection for DNA sampling, and thank Patricia Elda Rueda for plant genotyping. We thank Wojtek Pawlowski, Hong Ma and Yingxiang Wang for their RAD51 and γH2AX antibodies. We thank Shu-Chen Shen (Scientific Instrument Center, Academia Sinica) for help with super-resolution imaging. Special thanks to Thomas Boudier (Walter& Eliza Hall Institute, Australia) for his helps in image analysis using 3D Image Suite.
We thank Mathilde Grelon for her valuable suggestions for this manuscript. This research was supported by a grant from NIH GM48547 (to WZC), a research grant from UNAM PAPIIT . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018; IA201217 (to AR), a grant IOS-1339130 from the US National Science Foundation (to AB), and a grant (to CJRW) from Ministry of Science and Technology, Taiwan  . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018;  . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018;  (E-G) Transmission electron microscopy of synaptonemal complex spreads of wild-type (E), spo11-1-Δ03 (F) and spo11-1-Δ14 (G). Higher magnifications from F and G are shown on the right. In contrast to complete synapsis in wild-type, both mutants revealed promiscuous synapsis (f' and g') and curly structure (f'' and g''). Note that synapsis can occur within one axis, forming . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018; "fold-back" structures (arrows in f' and g'), and promiscuous synapsis can be also seen as unequal synapsis with an overhang in mutants (arrowheads in f'). Scale bars represent 1 μ m. (E-F) Super resolution images of DSY2 (red) and SPO11-1 (green) in wild-type meiocytes at late leptotene (E), early zygotene (F). SPO11-1 signals first appear mostly around axial elements, which are shown in a magnified image (e). Note that the wild type axis morphology at this stage is still curly. At the leptotene-zygotene transition (F), telomere bouquet forms (single bracket in F). Note that most of the SPO11-1 foci are close to axial elements at this stage (f).  Figure S1. Amplification of PhyC genes in wild-type (B73 and A344) and spo11-1 mutants.

Supporting Data
Both PhyC1 and PhyC2 genes were amplified using genomic DNA (gDNA) as template from B73 and A344 wild-types. 0P represents a negative control for PCR in absence of DNA template.

RT-PCR of PhyC gene expression and sequence analysis detected both PhyC1 and PhyC2
transcripts from cDNA of leaf (L) and anther (A) in wild-type (WT), in despite of one band. In contrast, sequencing results showed that only PhyC1 transcript was detected in mtm99-14 (14) and mtm00-03 (03) mutants.
. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
(I-L) spo11-1-1 meiocytes at metaphase I showed most univalents.   . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018;   Movie S1. A series of z-stack deconvolution microscopy images of DSY2 signals at early leptotene of wild-type.

Movie S2.
A series of z-stack deconvolution microscopy images of DSY2 signals at early leptotene of spo11-1-Δ03 mutant.
. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

Movie S4.
A series of z-stack deconvolution microscopy images with SPO11 (green) and DAPI-stained chromatin (red) at leptotene-zygotene transition of wild-type.

Movie S5.
A series of z-stack deconvolution microscopy images with SPO11 (green) and DAPI-stained chromatin (red) at pachytene of wild-type.
. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018;  . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.  a  u  d  a  t  F  ,  B  u  a  r  d  J  ,  G  r  e  y  C  ,  F  l  e  d  e  l  -A  l  o  n  A  ,  O  b  e  r  C  ,  P  r  z  e  w  o  r  s  k  i  M  ,  e  t  a  l  .  P  R  D  M  9  i  s  a  m  a  j  o  r  d  e  t  e  r  m  i  n  a  n  t  o  f  m  e  i  o  t  i  c  r  e  c  o  m  b  i  n  a  t  i  o  n  h  o  t  s  p  o  t  s  i  n  h  u  m  a  n  s  a  n  d  m  i  c  e  .  S  c  i  e  n  c  e  .  2  0  1  0  ;  3  2  7  (  5  9  6  7  )  :  8  3  6  -8  4  . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.  D  a  n  i  e  l  K  ,  R  o  i  g  I  ,  B  o  l  c  u  n  -F  i  l  a  s  E  ,  X  u  H  ,  B  o  o  n  s  a  n  a  y  V  ,  e  t  a  l  .  M  o  u  s  e  H  O  R  M  A  D  1  a  n  d  H  O  R  M  A  D  2  ,  t  w  o  c  o  n  s  e  r  v  e  d  m  e  i  o  t  i  c  c  h  r  o  m  o  s  o  m  a  l  p  r  o  t  e  i  n  s  ,  a  r  e  d  e  p  l  e  t  e  d  f  r  o  m  s  y  n  a  p  s  e  d  c  h  r  o  m  o  s  o  m  e  a  x  e  s  w  i  t  h  t  h  e  h  e  l  p  o  f  T  R  I  P  1  3  A  A  A  -A  T  P  a  . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018 ;   3  7  .  P  h  i  l  l  i  p  s  D  ,  N  i  b  a  u  C  ,  W  n  e  t  r  z  a  k  J  ,  J  e  n  k  i  n  s  G  .  H  i  g  h  r  e  s  o  l  u  t  i  o  n  a  n  a  l  y  s  i  s  o  f  m  e  i  o  t  i  c  c  h  r  o  m  o  s  o  m  e  s  t  r  u  c  t  u  r  e  a  n  d  b  e  h  a  v  i  o  u  r  i  n  b  a  r  l  e  y  (  H  o  r  d  e  u  m  v  u  l  g  a  r  e  L  .  ) . . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/489278 doi: bioRxiv preprint first posted online Dec. 6, 2018 ;   5  0  .  P  a  w  l  o  w  s  k  i  W  P  ,  G  o  l  u  b  o  v  s  k  a  y  a  I  N  ,  C  a  n  d  e  W  Z  .  A  l  t  e  r  e  d  n  u  c  l  e  a  r  d  i  s  t  r  i  b  u  t  i  o  n  o  f  r  e  c  o  m  b  i  n  a  t  i  o  n  p  r  o  t  e  i  n  R  A  D  5  1  i  n  m  a  i  z  e  m  u  t  a  n  t  s  s  u  g  g  e  s  t  s  t  h  e  i  n  v  o  l  v  e  m  e  n  t  o  f  R  A  D  5  1  i  n  m  e  i  o  t  i  c  h  o  m  o  l  o  g  y  r  e  c  o  g  n  i  t  i  o