ZmMTOPVIB Enables DNA Double-Strand Break Formation and Bipolar Spindle Assembly during Maize Meiosis1[OPEN]

The meiotic TopoVI B subunit (MTopVIB) plays an essential role in double-strand break formation in mouse (Mus musculus), Arabidopsis (Arabidopsis thaliana), and rice (Oryza sativa), and recent work reveals that rice MTopVIB also plays an unexpected role in meiotic bipolar spindle assembly, highlighting multiple functions of MTopVIB during rice meiosis. In this work, we characterized the meiotic TopVIB in maize (Zea mays; ZmMTOPVIB). The ZmmtopVIB mutant plants exhibited normal vegetative growth but male and female sterility. Meiotic double-strand break formation was abolished in mutant meiocytes. Despite normal assembly of axial elements, mutants showed severely affected synapsis and disrupted homologous pairing. Importantly, we showed that bipolar spindle assembly was also affected in ZmmtopVIB, resulting in triad and polyad formation. Overall, our results demonstrate that ZmMTOPVIB plays critical roles in double-strand break formation and homologous recombination. In addition, our results suggest that the function of MTOPVIB in bipolar spindle assembly is likely conserved across different monocots.

Meiotic homologous recombination is a crucial step for ensuring proper chromosome segregation and generating genetic diversity in eukaryotes (Keeney and Neale, 2006). During this process, induction of programmed DNA double-strand breaks (DSBs), which is catalyzed by an evolutionary conserved topoisomeraselike transesterase protein SPO11, initiates homologous recombination (Bergerat et al., 1997;Keeney et al., 1997). Through a transesterification reaction, the SPO11 dimer coordinately cleaves the double strands of a DNA molecule, forming an intermediate with one SPO11 molecule covalently linked with each 59 end of the cleaved DNA strands (Keeney and Kleckner, 1995). The SPO11 proteins are then released with the single-strand DNA oligos by the Mre11/Rad50/Xrs2 (MRX) complex and Sae2/CtIP (Neale et al., 2005;Symington and Gautier, 2011). The 59 ends are then resected to expose longer 39 single-strand DNA tails that are subsequently associated with two recombinases, RAD51 and DMC1, to form nucleoprotein filaments that promote homologous pairing by single-strand invasion into homologous chromosomes (Bishop et al., 1992;Shinohara et al., 1992;Cloud et al., 2012). The strand invasion results in a nascent DNA joint molecule called the displacement loop (D-loop;Hunter and Kleckner, 2001). Consequently, joint molecule intermediates can be processed into crossovers (COs) with reciprocal exchanges between homologous chromosomes or noncrossovers (Allers and Lichten, 2001;Baudat and de Massy, 2007;Grelon, 2016).
DNA topoisomerases are essential for regulating DNA topology, such as decatenating/relaxing superhelicity and untangling DNA during replication, transcription, and recombination (Corbett and Berger, 2003;Graille et al., 2008). DNA topoisomerases are classified into two families based on the cleavages they make on single-strand (type I) or double-strand DNA (type II), respectively (Champoux, 2001;Wang, 2002). The type II topoisomerases are predominant and can be further categorized into two subfamilies based on their structural similarity: type IIA and type IIB (Gadelle et al., 2014). Type IIA subfamily members, such as DNA gyrase and topoisomerase II enzymes, are found throughout eubacteria and eukaryotic organisms (Forterre and Gadelle, 2009). One type IIB subfamily member, topoisomerase VI (topo VI), is ubiquitous in archaea for decatenating DNA and is also found in plants, where it is needed for successful progression of the endoreduplication cycle (Bergerat et al., 1997;Hartung et al., 2002;Forterre et al., 2007). Topo VI functions in a heterotetramer complex comprised of two A (TopVIA) and two B (TopVIB) subunits (Corbett and Berger, 2003;Gadelle et al., 2014). About two decades ago, research showed that SPO11, which shares general sequence homology with the TopVIA, plays a critical role in DSB formation in all eukaryotes (Bergerat et al., 1997;Keeney et al., 1997). Interestingly, the analogs of the corresponding TopVIB and their critical roles in meiotic DSB formation have only recently been identified in mouse (Mus musculus; Robert et al., 2016a), Arabidopsis (Arabidopsis thaliana; Vrielynck et al., 2016) and rice (Oryza sativa; Fu et al., 2016;Xue et al., 2016).
In eukaryotes, faithful chromosome segregation during cell division is mediated by the spindle, i.e. a complex protein superstructure composed of microtubules and associated proteins (Heald et al., 1996;Compton, 2000;Wittmann et al., 2001;Xue et al., 2019). After nuclear envelope breakdown in plant meiosis, randomly polarized microtubules self-organize into bipolar spindles during metaphase I (Vernos and Karsenti, 1995;Heald et al., 1996;Gadde and Heald, 2004). A study in maize (Zea mays) has revealed that microtubules form around the condensed bivalents and then self-organize into a bipolar spindle (Chan and Cande, 1998). In maize mutants defective in chromosome synapsis and pairing, such as dsy1, dsy2, and afd1 meiocytes, spindle organization at metaphase I occurs normally, indicating that bipolar spindle formation is independent of paired kinetochores of bivalents (Chan and Cande, 1998). Therefore, the molecular pathways or regulatory factors responsible for the process of bipolar spindle assembly in maize remain elusive.
Very recently, the rice MTopVIB analog (OsMTOP-VIB) was found to play a key role in bipolar spindle assembly (Xue et al., 2019). The finding raises the critical question of whether this newly discovered role of OsMTOPVIB in meiosis is widely conserved across different plant species. In addition, a recent study detected few DSBs in the maize spo11-1 mutant (Ku et al., 2020), which suggested two possible scenarios: rare DSBs still occur in the absence of functional DSB machinery, or SPO11-2 is able to form a functional complex with MTIOVIB to execute DSBs. Here, we characterized functions of MTopVIB in maize, one of the best model organisms for cytogenetic study. We found that normal DSB formation was disrupted in Zmmtop-VIB mutant meiocytes, similar to observations in the maize spo11-1 mutant (Ku et al., 2020). A deficiency of DSBs led to defective homologous recombination and synaptonemal complex (SC) assembly, which consequently caused univalents at diakinesis. Moreover, bipolar spindle assembly was abnormal in ZmmtopVIB meiocytes, resulting in missegregated meiotic univalents which were aberrantly pulled by multipolarized spindles, yielding triads or polyads with micronuclei. Therefore, our results support the notion that MTOP-VIB displays a conserved function not only in DSB formation but also in bipolar spindle assembly among different monocots.

Identification of ZmMTOPVIB
To identify a putative MTOPVIB gene in maize, the fulllength amino acid sequence of Arabidopsis MTOPVIB was used as a query to search in the maize genome database (https://maizegdb.org/) using BLASTp analysis. We identified one candidate gene (Zm00001d014728) with the highest similarity to Arabidopsis MTOPVIB (At1G60460). By performing RACE, we obtained the 1,341-bp full-length ZmMTOPVIB complementary DNA (cDNA) sequence, which consists of 12 exons and 11 introns (Fig. 1A). The amino acid sequences of MTOPVIB from 10 different plant species obtained from the National Center for Biotechnology Information were subject to phylogenetic analysis. Two distinct clades of MTOPVIB homologs represented genetic divergence of monocots and dicots (Supplemental Fig. S1). In addition, alignment of MTOPVIB protein sequences from Arabidopsis, rice, and maize revealed that their MTOPVIB proteins are largely conserved, especially in three primary domains (GHKL, small, and transducer domains; Supplemental Fig. S2). We then investigated spatiotemporal expression patterns of ZmMTOPVIB by reverse transcription quantitative PCR (RT-qPCR) analyses and found that the gene was highly expressed in the developing tassel, moderately expressed in embryo, ear, and endosperm, and only weakly expressed in root, stem, and leaf (Supplemental Fig. S3).
To characterize functions of MTOPVIB in maize, we obtained one mutant line, UFMu-07260, from the Uni-formMu population (Liu et al., 2016), which has a Mutator inserted in the Zm00001d014728 gene. By PCR amplification and Sanger sequencing using the Mutator and ZmMTOPVIB-specific primers (Supplemental Table S1), we confirmed that the Mutator transposon was inserted into the first intron of ZmMTOPVIB (Fig. 1B). Although the insertion did not alter ZmMTOPVIB expression (Supplemental Fig. S3), it resulted in aberrant splicing, leading to two major splice variants. The longer transcript appeared more abundant relative to the shorter one, which was the same size as in the wild type (Fig. 1C). Sequence analyses revealed that the longer variant contains a fraction of Mutator transposon sequence, resulting in an in-frame premature stop codon (underlined tga; Fig. 1B). A second mutant allele (EMS4-0742ae) was obtained from the Maize EMS-induced Mutant Database (Lu et al., 2018). DNA sequence analysis of the EMS4-0742ae mutant confirmed a single base mutation from G to A at the splicing acceptor site of the eighth intron ( Fig. 1D), which was predicted to abort splicing of the eighth intron and result in an in-frame premature stop codon (underlined taa; Fig. 1D). RT-PCR analysis showed a longer transcript in the mutant, confirming retention of this intron (Fig. 1E).

Meiosis Is Disturbed in the ZmmtopVIB Mutants
Meiotic chromosome behavior in ZmmtopVIB mutants was analyzed by 4,6-diamidinophenylindole (DAPI) staining. In the wild type, long, thin, threadlike chromosomes were first observed at the leptotene stage ( Fig. 3A). Then, chromosomes rearranged next to a large and offset nucleolus and began to pair and synapse at the zygotene stage (Fig. 3B). During the pachytene stage, chromosomes completely synapsed to form thick chromosome threads (Fig. 3C). From the diplotene to diakinesis stage, chromosomes condensed into 10 short, rodlike bivalents distributed in the nucleus (Fig. 3D). At metaphase I, the 10 bivalents properly aligned on the equatorial plate (Fig. 3E), and equal numbers of chromosomes moved to the two opposite poles of the cell at anaphase I (Fig. 3F).
Chromosomal behaviors of ZmmtopVIB-1 and ZmmtopVIB-2 meiocytes were similar to the wild type in leptotene and zygotene (  ). Based on these obvious abnormalities in meiotic chromosome behavior, we concluded that the sterility of the ZmmtopVIB mutants was due to defective meiosis. Since both ZmmtopVIB-1 and ZmmtopVIB-2 exhibited similar meiotic phenotypes, all of our subsequent analyses were performed on ZmmtopVIB-1 as a representative mutant allele.

ZmMTOPVIB Is Essential for Normal Meiotic DSB Formation
Formation and repair of programmed DSBs during meiosis prophase I is an essential prerequisite for homologous recombination (de Massy, 2013). Despite important roles during meiosis, DSBs represent one of the most deleterious lesions, and such DNA damage rapidly induces phosphorylation of the histone variant H2AX at its S139 residue (gH2AX; Nakamura et al., 2010;Yuan et al., 2010). Consequently, the occurrence of gH2AX foci is routinely used as a biomarker to monitor DSB formation (Dickey et al., 2009;Löbrich et al., 2010). To investigate whether DSB formation is defective in maize ZmmtopVIB mutants, we used an antibody that specifically recognizes gH2AX for immunofluorescence analysis in wild-type and ZmmtopVIB-1 meiocytes. We found that gH2AX foci appeared as punctate-like signals scattered throughout the nuclei of wild-type meiocytes, reaching a peak at the early zygotene and decreasing at the late zygotene stage (Fig. 4, A-D and H). In contrast, very few gH2AX signals were detected at the early zygotene stage of ZmmtopVIB-1 meiocytes (Fig. 4, F and H), suggesting that normal DSB formation is defective in the ZmmtopVIB-1 mutant. Interestingly, at the late zygotene stage, we found a few clusters of gH2AX signals in ZmmtopVIB-1 meiocytes that were often associated with chromosome tangles (Fig. 4G). In magnified images, multiple chromosome axes appeared to become knotted around the gH2AX signals. On average, we found 19.9 gH2AX foci that often formed a few clusters in mutant cells (n 5 31; Fig. 4H).
To further investigate ectopic detection of gH2AX in mutants, we analyzed distribution of recombinases RAD51 and DMC1. While RAD51 is required for DNA repair in somatic and meiotic recombination, DMC1 functions specifically during meiotic recombination (Bishop et al., 1992;Cloud et al., 2012). In the wild type, both RAD51 and DMC1 manifested as numerous punctate foci distributed on chromosomes at the zygotene stage. At the late zygotene stage, their signals were significantly reduced (Fig. 5, A-D). In ZmmtopVIB-1 meiocytes, RAD51 and DMC1 were not detected in  the early zygotene stage (Fig. 5, E and G). Similar to gH2AX results, mutant meiocytes at the late zygotene stage displayed two to six clusters of RAD51 and DMC1 signals (Fig. 5, F and H). In magnified images, both signals were clearly decorated around these entangled chromosome axes (Fig. 5, I and J). Although DSBs were detected at a low level in mutant meiocytes, they were likely DNA damage associated with chromosomal entanglements. Given that no bivalent chromosomes were observed, these aberrant DSBs may not be repaired through the canonical meiotic recombination pathway. Nevertheless, these results suggest that ZmMTOPVIB is essential for normal meiotic DSB formation.

ZmMTOPVIB Is Critical for Homologous Pairing But Not Required for Telomere Bouquet Formation
To investigate whether defective ZmMTOPVIB disrupts homologous chromosome pairing, we performed fluorescence in situ hybridization (FISH) using the 5S ribosomal DNA (rDNA) probe in the wild type and the ZmmtopVIB-1 mutant. The 5S rDNA is a tandem repeat sequence located on the long arm of chromosome 2 that is routinely used to evaluate chromosome pairing and segregation in maize . In wild-type meiocytes, two 5S signals gradually paired with each other during the zygotene stage (Fig. 6, A and B). At the pachytene stage, the paired 5S signal was observed in all cells examined (n 5 48; Fig. 6C). In contrast, two separate 5S rDNA signals were consistently detected in ZmmtopVIB-1 meiocytes (Fig. 6, F and G), indicating that homologous chromosome pairing is defective in the ZmmtopVIB-1 mutant.
The telomere bouquet is an evolutionarily conserved chromosome arrangement that clusters telomeres in a small region on the nuclear envelope (Tomita and Cooper, 2007;Moiseeva et al., 2017). This specialized structure is thought to promote initiation of homologous pairing during early prophase I (Loidl, 1990;Scherthan, 2001;Harper et al., 2004). To explore whether defective ZmmtopVIB affects telomere bouquet formation, we conducted FISH analysis using the pAtT4 probe (Richards and Ausubel, 1988;Prieto et al., 2004) in wildtype and ZmmtopVIB-1 meiocytes. In wild-type (n 5 44;    6, A and B) and ZmmtopVIB-1 (n 5 51; Fig. 6F) meiocytes at the early zygotene stage, telomere signals were clustered and attached to the nuclear envelope, indicating that ZmMTOPVIB is not required for telomere bouquet formation. As mutant meiocytes did not show a normal pachytene stage, telomere signals become diffused at the late zygotene (Fig. 6G).
The observation of gH2AX signals in ZmmtopVIB mutants was not expected. Although homologous pairing assessed by 5S rDNA signals suggested that pairing is defective, we sought to analyze more loci in the genome. Therefore, we adapted the recently developed chromosome painting method in meiosis (Albert et al., 2019). By using chromosome 3-and chromosome 8specific probes, wild-type meiocytes clearly showed 10 bivalent chromosomes at diakinesis (Fig. 6D). Bivalents labeled with the chromosome-specific probes were correctly separated at anaphase I (Fig. 6E). In contrast, ZmmtopVIB mutants showed a complete failure of bivalent formation by chromosome painting (Fig. 6, H and I).

ZmmtopVIB-1 Meiocytes
The SC is a meiosis-specific chromosomal structure that connects homologous chromosomes by a transverse filament to promote efficient CO formation (Page and Hawley, 2004;Argunhan et al., 2017). To investigate Figure 6. ZmMTOVIB is essential for homologous pairing but not for telomere bouquet clustering. A to C, FISH in wild-type meiocytes using 5s rDNA (red) and telomere (green) probes at different stages. D and E, Chromosome (Ch) painting in wild-type meiocytes using chromosome 3-(red) and chromosome 8-specific (green) probes. F and G, FISH in ZmmtopVIB-1 meiocytes using 5s rDNA (red) and telomere (green) probes at different stages. White arrows point out the unpaired 5s foci in ZmmtopVIB-1. H and I, Chromosome painting in ZmmtopVIB-1 meiocytes using chromosome 3-(red) and chromosome 8-specific (green) probes. Scale bars 5 5 mm.
the pattern of SC localization in wild-type and ZmmtopVIB-1 mutant meiocytes, we performed immunostaining using AFD1, ASY1, and ZYP1 antibodies. Maize AFD1 is homologous to Arabidopsis and rice REC8 (Cai et al., 2003;Shao et al., 2011). It is a vital component of the cohesion complex associated with axial and lateral elements and is required for axial element (AE) elongation (Golubovskaya et al., 2006). AFD1 signals appeared as threads along entire chromosomes of wild-type meiocytes at the zygotene stage (n 5 21; Fig. 7A). AFD1 signals in ZmmtopVIB-1 meiocytes was consistent with that of the wild type (n 5 24; Fig. 7B), indicating that ZmMTOPVIB is not required for cohesion complex assembly.
Maize ASY1, a homolog of Arabidopsis ASY1 and rice PAIR2, is a basic element of AE, and it is essential for SC assembly and homologous recombination (Armstrong et al., 2002;Golubovskaya et al., 2011). In wild-type meiocytes, ASY1 loading manifested as linear signals along entire chromosomes in the zygotene stage (n 5 22; Fig. 7C). Again, ASY1 signals in ZmmtopVIB-1 meiocytes were very similar to the pattern observed for the wild type (n 5 20; Fig. 7D), confirming that ZmMTOPVIB is not required for AE installation.
Maize ZYP1, a homolog of Arabidopsis ZYP1 and rice ZEP1, is the central element (CE) of the SC and is assembled between AEs to regulate chromosome synapsis and CO formation (Higgins et al., 2005;Wang et al., 2010;Barakate et al., 2014). In wild-type meiocytes, we observed elongated filaments of ZYP1 signals along the entire length of synapsed chromosomes at thte pachytene stage (n 5 25; Fig. 7E). In contrast, only short branched or punctate ZYP1 signals were observed in ZmmtopVIB-1 meiocytes at the same stage (n 5 27; Fig. 7F), supporting the finding that ZmMTOPVIB is crucial for ZYP1 loading. Taken together, these results indicate that ZmMTOPVIB is not required for AE installation, but it is indispensable for CE assembly during maize meiosis.

ZmMTOPVIB Is Required for Meiotic Bipolar Spindle Assembly in Maize
The recently identified function of OsMTOPVIB in regulating meiotic spindle assembly in rice prompted us to investigate whether ZmMTOPVIB has the same role in maize meiosis. To do so, we performed immunostaining using a-tubulin antibody in wild-type and ZmmtopVIB-1 meiocytes. a-Tubulin heterodimerizes and polymerizes with b-tubulin to form microtubule walls (Hunter and Kleckner, 2001;Blume et al., 2009;Gunning et al., 2015;Higgins et al., 2016). Microtubule filaments in wild-type meiocytes gradually extended and attached to chromosomes during diakinesis (n 5 13; Fig. 8A). At metaphase I, spindle fibers linked to the kinetochores and were arranged into two arrays, forming a canonical bipolar spindle structure (n 5 25; Fig. 8B). At anaphase I, as the spindles extended, equal numbers of chromosomes were pulled by the bipolar spindles toward the opposite poles of the cell (n 5 23; Fig. 8C). Finally, dyads formed, and the remaining spindles stayed at the equatorial plate (n 5 18; Fig. 8D). In contrast, we consistently observed abnormal spindle structures in the ZmmtopVIB-1 mutant at different stages of meiosis. In diakinesis, extension and aggregation of microtubule filaments was incomplete, with only one polar spindle forming, so that the opposite cell pole lacked a polar spindle (n 5 11; Fig. 8, E and I). At metaphase I, spindle fibers of mutant meiocytes became entangled and distorted, lacking the typical bipolar spindle structure upon separation (n 5 12; Fig. 8, F To ascertain the outcomes of these abnormal spindle structures, we examined spore products in wild-type and ZmmtopVIB-1 lines. In the wild type, we consistently observed symmetric dyads at telophase I (n 5 55; Fig. 9A) and tetrads at telophase II (n 5 52; Fig. 9E). Similarly, we also observed dyads at telophase I (54.13%, n 5 59 ; Fig. 9B) and tetrads at telophase II (36.76%, n 536; Fig. 9F) in ZmmtopVIB-1. In contrast, we frequently detected triads at telophase I (25.69%; n 5 28; Fig. 9C) and polyads at telophase I and II (63.24%; n 5 62; Fig. 9, D, G, and H) in ZmmtopVIB-1. Taken together, these results demonstrate that ZmMTOPVIB plays a critical role in proper bipolar spindle assembly during maize meiosis.

Evolutionary Conservation of ZmMTOPVIB Function in Meiotic DSB Formation
We have shown that disruption of ZmMTOPVIB causes defects in the formation of meiotic DSBs and results in severe abnormalities of DSB-induced events, such as homologous synapsis, meiotic recombination, and chromosome segregation. Although we detected few gH2AX signals in the late zygotene stage, often associated with entangled chromosomes in mutant meiocytes, these sites are unlikely to be canonical meiotic DSBs. One possible scenario is that DNA damage resulted from chromosome knotting, which may be more frequent in a large genome with long chromosomes in the absence of recombination initiation. In support of this possibility, a similar observation was reported in maize spo11-1 mutants (Ku et al., 2020). Nevertheless, our results demonstrate that ZmMTOP-VIB is essential for normal meiotic DSB formation in maize, consistent with studies in other organisms, including mouse (Robert et al., 2016a), Arabidopsis (Vrielynck et al., 2016), and rice (Fu et al., 2016;Xue et al., 2016), strengthening the notion that the role of TopoVIB in meiotic DSB formation is evolutionarily conserved among plants and mammals.
TopoVI is a heterotetramer comprised of two A subunits and two B subunits (Corbett and Berger, 2003;Gadelle et al., 2014). To generate DSBs on DNA, the A 2 dimers catalyze the transesterification reaction for DNA cleavage (Bergerat et al., 1997), whereas the B 2 subunits are responsible for A 2 conformation and DNA binding (Dutta and Inouye, 2000;Corbett et al., 2007;Graille et al., 2008). Although the basic setup of the TopoVI complex seems conserved among organisms, the arrangement and operational mode of each component may vary. For instance, in mouse, the sole SPO11 gene produces two major SPO11 isoforms, i.e. a short SPO11a and a longer SPO11b, by alternative splicing (Bellani et al., 2010). SPO11b interacts with TopoVIB to form a symmetrical heterotetramer for creating meiotic DSBs (Robert et al., 2016b). In contrast, all land plants encode at least two SPO11 proteins, referred to as SPO11-1 and SPO11-2, both of which are required for DSB formation (Sprink and Hartung, 2014). Accordingly, the plant TopoVI complex is predicted to exist as an asymmetrical heterotetramer (Graille et al., 2008). Interestingly, Arabidopsis MTOPVIB not only acts as an essential component of the TopoVI complex but also plays a critical role in mediating the formation of Figure 8. Meiotic spindle assembly process in the wild type (A-D) and ZmmtopVIB-1 (E-L) in diakinesis (A, E, and I), metaphase I (B, F, and J), anaphase I (C, G, and K), and telophase I (D, H, and L) stages. Chromosomes were marked with DAPI (blue) and microtubules with a-tubulin antibody (green). Red arrows point out the abnormal spindles and equatorial plates. Scale bars 5 10 mm. SPO11-1/SPO11-2 heterodimers, though not of SPO11-1 or SPO11-2 homodimers (Vrielynck et al., 2016;Jing et al., 2019b). Hence, this function of Arabidopsis MTOPVIB as a linker to promote the assembly of SPO11 dimers may not necessarily be displayed by its counterpart in mouse. These results highlight that analogs of TopoVIB involved in meiotic DSB formation are evolutionarily conserved but subject to variation among different organisms.
Although we detected low levels of gH2AX, RAD51, and DMC1 signals in the ZmMTOPVIB mutant, these signals appear to aggregate abnormally around chromosomal entangles. These knotted chromosome axes may reflect difficulties of pairing due to lack of normal DSB formation. Perhaps these entanglements may trigger TOPII activity for resolution, which ectopically creates chromosomal breaks. The recent study of Arabidopsis TOPII and its role in unknotting entangled chromosomes also supports this possibility (Martinez-Garcia et al., 2018). In addition, a similar phenotype of aberrant DSB formation was also reported in the maize spo11-1 mutant (Ku et al., 2020), which may imply that species with long meiotic chromosomes endure a higher risk of chromosome damage when DSBs fail to initiate. However, given the absence of normal bivalent chromosomes and chromosome fragmentation in ZmMTOPVIB mutant cells, these aberrant DSBs were likely repaired but did not give rise to meiotic crossovers.

ZmTOPVIB Is Involved in Meiotic Bipolar Spindle Assembly
Proper morphogenesis and orientation of microtubulebased spindles are critical processes for ensuring the fidelity of chromosome segregation and cell division in eukaryotes (Gadde and Heald, 2004;Zhang and Dawe, 2011). In both monocot and dicot plants, construction of a meiotic bipolar spindle occurs by converting multipolar into bipolar spindle poles Xue et al., 2019). However, the mechanism underlying meiotic bipolar spindle assembly remains unclear, with only a few contributory genes reported to date. Arabidopsis ATK1 and ATK5, which encode kinesins, are motor proteins that typically advance along microtubules and presumably generate the force for microtubule movement, thereby affecting spindle organization (Chen et al., 2002;Ambrose and Cyr, 2007). PRD2, an essential gene for meiotic DSB formation in Arabidopsis, was originally Figure 9. Meiotic products in the wild type and ZmmtopVIB-1. A to H, Shown are meiotic cell division in the wild type (A and E), bipolar division of ZmmtopVIB-1 (B), producing tetrads (F); tripolar division of ZmmtopVIB-1 (C), producing pentads/ hexads (G), multipolar division of ZmmtopVIB-1 (D), and producing polyads (H). Red arrows point out the micronucleus and abnormal additional spores. Scale bars 5 10 mm. I, Percentage of different meiotic products (dyad, triad, tetrad, and polyad).The left two columns show meiotic products in telophase I, and the right two columns meiotic products in telophase II.
named Multipolar Spindle 1 (MPS1) based on the unequal bipolar or multipolar spindles present in mps1 mutant meiocytes (Jiang et al., 2009). Intriguingly, OsMTOPVIB interacts with OsPRD2 in rice, indicating that these proteins may function in the same pathway, though the function of OsPRD2 has not yet been characterized (Xue et al., 2019). Since both PRD2 and OsMTOPVIB are essential for DSB formation, it is conceivable that the processes of DSB formation and bipolar spindle assembly may be coordinated. However, this speculation was proven untrue upon discovery that bipolar spindles at metaphase I occur normally in three rice mutants exhibiting defects in DSB formation, i.e. the pair2, Osspo11-1, and crc1 mutants (Xue et al., 2019). Given these results, it is evident that meiotic spindle assembly is independent of DSB formation. Additionally, bipolar spindle assembly also seems to be uncoupled from homologous pairing, since it was not perturbed in the dsy1, dsy2, and afd1 mutants, three maize mutants defective in homologous pairing (Chan and Cande, 1998). Therefore, the molecular mechanism that integrates these homologous recombinationrelated genes in meiotic spindle assembly remains to be resolved.
As reported for the OsmtopVIB mutant, we observed a substantial proportion of triads at telophase I and polyads at telophase II in ZmmtopVIB meiocytes, demonstrating that maize ZmMTOPVIB is also required for meiotic bipolar spindle assembly. This finding confirms functional conservation of the MTOPVIB role in bipolar spindle assembly between rice and maize. However, we observed fewer abnormal triads or polyads in ZmmtopVIB than were found in the Osm-topVIB mutant line. That discrepancy may be due to differential allelic effects on phenotypes. The two mutant alleles we considered in this study exhibited defective transcript splicing, hypothetically resulting in production of truncated protein. Although the complete abortion in DSB formation indicates that two alleles are at least severe hypomorphs, we cannot rule out the possibility that neither of these ZmmtopVIB mutant alleles was completely null for the specific role in bipolar spindle assembly, unlike the previously studied OsmtopVIB mutant. Therefore, it would be worth investigating spindle defects in other ZmmtopVIB mutants in future studies.
Overall, our study demonstrates that ZmMTOPVIB is essential for meiotic DSB formation and that it plays a critical role in bipolar spindle assembly in maize. Our findings shed light on the evolutionary conservation of the dual functions of MTOPVIB in meiosis, though the mechanisms by which it plays a moonlighting role in bipolar spindle assembly await further investigation.

Plant Material and Genotyping
The maize (Zea mays) UFMu-07260 mutant line (ZmmtopVIB-1) in the W22 inbred background was obtained from the UniformMu stock center of MaizeGDB (https://maizegdb.org/; McCarty et al., 2013). Another mutant allele, EMS4-0742ae (ZmmtopVIB-2) in the B73 inbred background, was obtained from the Maize EMS Induced Mutant Database (http://www. elabcaas.cn/memd/; Lu et al., 2018). All plants were cultivated and fertilized under normal field conditions during the growing season or in a growth chamber (16 h light at 28°C, 8 h dark at 22°C, 60% to 70% humidity). Maize genomic DNA was extracted using a method previously described (Li et al., 2013). Primers used for genotyping and sequencing of the two mutant alleles are listed in Supplemental Table S1.

Pollen Viability
Pollen viability was assessed using the Alexander staining method (Alexander, 1969;Johnson-Brousseau and McCormick, 2004). Mature pollen grains were dissected out of anthers from the wild type and ZmmtopVIB mutants during the pollination stage and then stained with 10% Alexander solution. Images of stained pollen grains were taken using a Leica EZ4 HD stereo microscope equipped with a Leica DM2000 LED illumination system.

RT-qPCR Analysis
Total RNA was isolated from root, stem, leaf, developing embryo, endosperm, young tassel, and young ear using a TRNzol-A 1 Kit Reagent (TIANGEN) according to the manufacturer's instructions. Reverse transcription was performed using a PrimeScript II first strand cDNA Synthesis Kit (TaKaRa) with Oligo-T primers to obtain cDNA. Quantitative PCR was conducted with a 7500 Fast Real-Time PCR System (Applied Biosystems) using SYBR Green Master Mix (TaKaRa). All reactions were performed with three biological replicates and technical repeats. Gene-specific primers and reference gene (Ubiquitin) primers for internal control are listed in Supplemental Table S1.

Meiotic Chromosome Preparation and DAPI Staining
Young tassels were fixed in Carnoy's solution (ethanol:glacial acetic acid [3:1]) for 1 d at room temperature and then stored in 70% (v/v) ethanol at 4°C. Anthers were dissected in 45% (v/v) acetic acid solution. Meiocytes were squeezed from anthers and squashed onto slides using coverslips. Slides were frozen in liquid nitrogen and the coverslips were removed immediately. After serial dehydration in 70%, 90%, and 100% (v/v) ethanol, the air-dried slides were stained and mounted with DAPI in Vectashield antifade medium (Vector Laboratories).

FISH and Chromosome Painting
Chromosome spreads were prepared by the method described previously . Three repetitive DNA probes were used, including the pTa794 clone containing 5S rDNA repeats (Li and Arumuganathan, 2001), the pAtT4 clone containing telomere-specific repeats (Richards and Ausubel, 1988), and cy5-conjugated 180-bp knob oligonucleotides. The rDNA and telomere probes were labeled by the Nick Translation Kit (Roche). The chromosome 3 painting probe was labeled with ATTO-550 as previously described (Albert et al., 2019). Slides were counterstained using DAPI in antifade mounting medium (Vector Laboratories). Chromosome images were captured under a Ci-S-FL fluorescence microscope (Nikon) equipped with a DS-Qi2 microscopy camera (Nikon) or under a Delta Vision ELITE system (GE Healthcare) equipped with an Olympus IX71 microscope.

Immunofluorescence Assay
Immunofluorescence was performed as described previously (Pawlowski et al., 2003), with minor modifications. After being dissected and permeabilized in 13 buffer A solution with 4% (w/v) paraformaldehyde for 30 min at room temperature, fresh young anthers were washed twice in 13 buffer A at room temperature and then stored in 13 buffer A at 4°C. Meiocytes were squeezed from anthers and squashed onto slides. After freezing in liquid nitrogen, coverslips were removed immediately. The meiocytes were incubated in blocking buffer diluted with primary antibodies for 1 h in a 37°C humidity chamber, then washed three times in 13 phosphate-buffered saline. Goat anti-rabbit antibodies conjugated with Fluor 555 diluted in blocking buffer were added to the slides. After incubation at 37°C for 1 h, the slides were washed three times in 13 phosphate-buffered saline. Finally, cells were counterstained with DAPI in antifade mounting medium (Vector Laboratories). The antibodies against ASY1, ZYP1, and gH2AX were prepared as described previously (Jing et al., 2019a). Antibodies against AFD1, RAD51, and DMC1 were gifts from collaborators. All primary and secondary antibodies were diluted at 1:100. Images of meiocytes were observed and captured using a Ci-S-FL microscope (Nikon) equipped with a DS-Qi2 microscopy camera (Nikon). Two-dimensional projected images were generated using NIS-Elements software. Further image processing was conducted using ImageJ software (https://imagej.nih.gov/ij/ index.html).

Supplemental Data
The following supplemental materials are available.
Supplemental Figure S1. Phylogenetic analysis of MTOPVIB homologs in different plant species.
Supplemental Figure S2. Multiple sequence alignment analysis of MTOP-VIB proteins in maize, Arabidopsis, and rice.
Supplemental Table S1. Primers used in this study.