The role of DNA topoisomerase 1α (AtTOP1α) in regulating arabidopsis meiotic recombination and chromosome segregation

Plant materials, crosses, and growth conditions

A. thaliana Columbia (Col-0) plants were used in this study. The top1α1 had a T-DNA insertion in the eighth intron (Takahashi et al., 2002), and was obtained from the Taku Takahashi lab. The atm-2 mutant (SALK_006953) was obtained from the Salk Institute for Biological Studies. The double mutant was generated by crossing top1α1 and atm-2 mutants. Plants were maintained at ~22 °C and 60% humidity with light cycles of 16 h light/8 h dark.

PCR genotyping, semi-quantitative PCR and qRT-PCR

Plant genotyping was performed using PCR with specific primers detailed in Table S1. These primers included top1α_LP, RP and LB; atm_LP, RP and SALK LBb1.3. All plant lines were in the Columbia (Col-0) genetic background. RNA extraction from seedlings was achieved using TRIzol reagent (Invitrogen, Waltham, MA, USA). Subsequently, first-strand cDNA synthesis was carried out from 3 μg of total RNA using 200 units (U) of M-MLV reverse transcriptase (Invitrogen, Waltham, MA, USA). qRT-PCR was performed using an optical 96-well plate in an ABI Step one plus PCR system (Applied Biosystems, Waltham, MA, USA) using SYBR Green Master Mix (Roche, Basel, Switzerland). ACTIN8 served as the reference gene for the normalization of gene expression.

Pollen grains, mature anther staining, and dissected tetrads

Pollen viability and mature anther phenotypes were assessed using Alexander stain according to the Gu method (Liu et al., 2023). Briefly, flower buds at stage 12 of anther development were fixed in 1 mL of Carnoy’s solution (6:3:1 ethanol:chloroform:acetic acid, v/v) for a minimum of 2 h. Following fixation, the Carnoy’s solution was removed, and the flower buds were dissected under a dissecting microscope. To differentiate viable from non-viable pollen grains, individual anthers were incubated with Alexander stain for 30 min at room temperature (25 °C). Subsequently, the anthers were mounted on a slide and visualized using an Olympus BX53F microscope (Olympus, Tokyo, Japan). This staining technique enabled the discrimination of viable pollen grains (magenta-red) from non-viable ones (blue or green) (Chang et al., 2014; Mao-Sen et al., 2021).

Iodine pollen starch test

Pollen viability was assessed using a starch staining technique as described by Chang et al. (2021). Inflorescences were fixed and stored in 70% (v/v) ethanol. Flower buds were dissected under a dissecting microscope for anther isolation. Six anthers per flower were crushed on a microscope slide in a droplet of 1% (w/v) iodine-potassium iodide (I2/KI) solution. Excess solution was carefully withdrawn using forceps. The preparation was then observed under a light microscope at room temperature (25 °C) after coverslip placement.

Separation and detection of tetrads

Inflorescences with unopened flowers were fixed with Carnoy’s fixative (3:1 100% ethanol, glacial acetic acid) overnight at 25 °C. Older buds outside the inflorescence containing obvious yellow anthers were removed, while sepals were left intact. The inflorescence was washed three times with water after preparation. Immature buds were dissected using a fine needle on a glass slide under a binocular microscope. Immature anthers were isolated from other floral organs in two drops of water. The anthers were then squeezed with sharp forceps to release the tetrads. The slide was air-dried, and tetrads were stained for 2–3 min with 0.01% basic fuchsine (1:3 3.0 g basic fuchsine in 100 mL 95% ethanol, 5% aqueous phenol). The slide was covered with a coverslip, and the red-stained tetrads were observed under a light microscope following previously recommended procedures (Prusicki et al., 2021).

Chromosome distribution for meiotic chromosome detection with DAPI

The chromosome spread was determined by applying the protocol described previously (Prusicki et al., 2021; Wang et al., 2014). The inflorescences with unopened flowers were fixed with Carnoy’s solution at room temperature overnight. The buds rinsed three times every 10 min with 10 mM citrate solution. Then, old buds were removed containing yellow anthers that could be seen before adding the digestion-by-digestion cocktail (0.3% pectolyase, 0.3% cellulase, and 0.5% cytohelicase in 10 mM citrate buffer, pH 4.5) at 37 °C for 80 min. Then, we washed the set buds five times, every 5 min, with 10 mM citrate buffer, after stopping digestion on ice for 5 min. Following that, we placed a bud on the slide to dissect it, obtain anthers, and remove the rest of the flower. Anthers were mashed with sharp forceps in a drop of 60% acetic acid. Next, we moved the slide to a heat block, incubated it at 45 °C for 30 s, and added another drop for another 30 s. Then, 20 µL of freezing Carnoy’s fixative was added to a center sample to spread the meiotic chromosomes on the slide, which were left to air dry for approximately 5 min. We added 5 μL of DAPI to the slide, covered with a coverslip, and observed the chromosomes under a fluorescent microscope.

Fluorescence In Situ Hybridization (FISH)

Chromosomes were separated according to Wang et al. (2014). Thereafter, 100 μL of 70% formamide was added to each dry slide dissolved in 2 × SSC buffer; then, the slide was covered with a parafilm piece and incubated in an oven at 80–90 °C for 5 min. The slides were dehydrated by treating with cold alcoholic dilutions at −20 °C (70%, 80%, 90%, and 100% ethanol) for 5 min for each dilution, and then left the slides to air dry. Following that, 10 μL of solution was mixed by combining 2 μL of the labeled probe (centromere or/and telomere or/and 45s rDNA probe, designed as shown in Table S2) with 8 μL the hybridization cocktail for each slide. We poured 10 μL mixture for each slide and covered with a coverslip; slides were incubated at 85 °C for 5 min to denature the mixture and then cooled directly on ice for 5 min. The slices were incubated in a slide box with high humidity to prevent the slices from drying out at 37 °C overnight while avoiding bright light. The covers were removed in a 2 × SSC buffer and then washed in 2 × SSC buffer three times for 15 min each time, and then one time with 1 × PBS buffer for 5 min. The slides were air-dried, 10 μL DAPI (4, 6-diamidino-2-phenylindole) was applied on each slide, and we covered with a coverslip. The slides were observed under a fluorescent microscope.

Statistical analysis

All experiments were performed at least three times. For statistical comparisons, we used Student’s t-test. Data shown are averages ± SD.

The absence of TOP1α resulted in reduced fertility and impaired the process of meiosis

To investigate how the absence of functional $AtTOP1\alpha$ affects pollen viability and anther starch accumulation, we used Alexander Red staining to assess pollen viability and PCR analysis to confirm the genotype of top1α1 mutants. The homozygous mutants developed normally, but with early flowering and differences in leaf shape and internodal distance (Takahashi et al., 2002). Since the T-DNA insertion disrupts $AtTOP1\alpha$ gene expression, a continuous mRNA transcript cannot be generated across the insertion site. Therefore, the expression level of AtTOP1α cannot be determined using the presented semi-quantitative PCR (Fig. 1F) and qPCR (Fig. 1G) data. Our results revealed that the wild-type (WT) anthers contained significantly more pollen grains compared to top1α1 anthers (Figs. 1A and 1B, respectively). We also examined the tetrad stages, while the WT tetrads had four microspores (Fig. 1C), tetrads contained polyads with small microspores, suggesting the occurrence of a meiotic abnormality. A total of 50 observations were conducted, revealing that 6% of the observed tetrads had abnormalities (Figs. 1D and 1E). We also investigated pollen fertility using I2/KI staining and observed many immature pollen grains in the mutant; some of the top1α1 pollen grains were larger than wild-type pollen (Fig. S1).

The disruption of TOP1α influenced the behavior of chromosomes during prophase

To assess chromosomal behavior during male meiosis in WT and top1α1 mutants, we used DAPI staining. DAPI staining revealed largely normal meiosis in the mutants compared to WT, with minor variations in some stages (Fig. 2). Both WT and top1α1 meiotic chromosomes condensed at leptotene and displayed generally identical morphologies (Figs. 2A and 2A1, respectively). However, in top1α1 at zygotene (n = 465/589) and pachytene (n = 628/807), a chromosome displayed bright large blocks of heterochromatin (Figs. 2B1 and 2C1, respectively). WT meiocytes (n = 259) had compact homologs at diplotene (Fig. 2D), but the top1α1 (n = 289/342) chromosomes were fragile, with thinner regions than usual (Fig. 2D1). In diakinesis, the WT (n = 196) showed five complete bivalents (Fig. 2E), while some entanglements existed between the bivalents in top1α1 (n = 178/211) (Fig. 2E1). At metaphase I, the WT bivalents (n = 192) were regular and lined up (Fig. 2F), while the top1α1 bivalents (n = 99/156) were often thin with a few crossovers and two or more top1α1 chromosomes frequently linked together (Fig. 2F1). After meiotic II division, chromosome crowding was observed in top1α1 at anaphase I and telophase I (Figs. 2G1 and 2H1, respectively). The presence of stacking during meiosis I was consistent with the lack of $AtTOP1\alpha$ function, which plays a critical role in developing appropriate DNA structure and topology, fixing kinks, torsions, and dealing with tight-twisted intermediates produced during replication.

TOP1α is required to untangle chromosomes at the centromere regions

The centromere is a region on each linear chromosome essential for ensuring proper chromosome segregation during cell division. To test whether the absence of the TOP1α function affects the disentanglement at the chromosome level during meiosis I, we performed FISH using a centromere probe on top1α1 meiotic cells to examine chromosome pairing and synapsis, to see whether top1α1 bivalents occurred between homologs or non-homologs. FISH analyses with a centromere probe exhibited a similar number of signals in WT and top1α1 at leptotene, with wild-type cells displayed 10 centromere foci, while top1α1 cells had 8–10 foci (Figs. 3A and 3B, respectively).

At zygotene and pachytene, Col-0 WT cells had 10 (n = 85) and 5 (n = 120) centromeric foci, respectively (Figs. 3C and 3E); in contrast, there were fewer and larger foci in top1α1 at zygotene, ranging between 6 (~16%, 28/175), 5 (~26%, 45/175), 4 (~23%, 40/175), 3 (~23%, 40/175), 2 (~9%, 15/175), and 1 (~3%, 5/175) (Fig. 3D). Several top1α1 cells at the zygotene stage also displayed clustered foci in one location compared to WT (Fig. S2).

In top1α1, the number of chromosome signals decreased at the pachytene stage. The number of foci ranged from 5 (~11.5%, 30/260), 4 (~19%, 50/260), 3 (~52%, 135/260), 2 (~13.5, 35/260), to 1 (~4%, 10/260), with a large signal (Fig. 3F), which was confirmed by multiple cell analysis of top1α1 at the pachytene stage (Fig. S3).

At diakinesis, the WT had five pairs of homologous chromosomes and five centromeric foci (Fig. 3G). In contrast, the top1α1 chromosomes were grouped and overlapping, with signals concentrated in two locations (Fig. 3H). At metaphase I, in WT, five pairs of centromere foci on five bivalents were positioned on a single line (Fig. 3I), while in top1α1 we observed that, chromosomes were in abnormal alignment, and there was a tangle between non-homologous chromosomes but also 10 foci (Fig. 3J), revealing a non-homologous association. Results for WT and top1α1 cells from anaphase I to telophase II were similar (Fig. 3). From these observations, we suggest that top1α1 had abnormalities in pairing and synapsis around the centromeres.

In many organisms, telomeres are arranged as “a bouquet” on the nuclear envelope to promote pairing and synapsis. While Arabidopsis does not have a traditional bouquet, it is believed that the same outcome is achieved by clumping telomeres around the nucleolus before leptotene. We observed this “clumping” pattern in both WT and top1α1 telomere foci at leptotene (Figs. S4A and S4B, respectively), at zygotene (Figs. S4C and S4D, respectively), and at pachytene (Figs. S4E and S4F, respectively), and we found WT and top1α1 both had similar numbers of signals (9–10, 10 and 10 telomere foci, respectively).

Disruption of TOP1α affected 45s DNA localization in A. thaliana

We conducted FISH analysis targeting 45S rDNA to examine chromosome pairing in top1α1 plants compared to WT. The 45s rDNA fociin WT and top1α1 were identical at leptotene and zygotene (Figs. 4A, 4C and 4B, 4D, respectively). At pachytene, the WT had a single intense signal (Fig. 4E), while the majority of the top1α1 samples had two separate signals (Fig. 4F), which indicates partial separation of the 45S rDNA regions. This result shows that homologous chromosomes in top1α1 were incorrectly paired and synapsed easily at the 45S rDNA region. We also observed the presence of signals on three bivalents in top1α1 at diakinesis and metaphase I (Figs. 4H and 4J, respectively) and more images at metaphase I (Fig. S5), to illustrate these pairing abnormalities in top1α1 cells at metaphase I.

TOP1α promotes crossover in A. thaliana

Visualizing of chiasma during diplotene and diakinesis in Arabidopsis thaliana is hindered by the difficulty in differentiating between true chiasmata and non-recombinant twists of homologous chromosomes. Additionally, the close association of nucleolus organizer regions (NORs) with the nucleolus in the short arms of chromosomes 2 and 4 impedes chiasma identification in these regions. Therefore, despite the condensed state of chromosomes at metaphase I, this stage offers a more suitable platform for chiasma scoring in Arabidopsis. We investigated and documented chiasma frequency in wild-type (WT) and top1α1 mutant pollen mother cells (PMCs) at metaphase I using fluorescence in situ hybridization (FISH) labeling with a centromeric probe, following established protocols. Analysis revealed a consistent pattern of five bivalents formed by the five chromosome pairs in all examined cells. These bivalents could be categorized as rods or rings. Rods displayed chiasmata in only one chromosome arm, while rings exhibited chiasmata in both arms. Figure 5 demonstrates a significant difference in chiasma number between WT and top1α1. The mutant exhibited an increased frequency of cells containing 5, 6, or 7 chiasmata compared to WT. Conversely, WT had a higher proportion of cells containing more than seven chiasmata. This suggests that functional TOP1α does not influence crossovers in Arabidopsis. Based on these findings, we conclude that TOP1α promotes crossover events in Arabidopsis.

TOP1α acts synchronously with ATM during DNA repair

The atm mutant is known to have defects in meiosis and produces fewer numbers of seeds. While the pachytene stage in atm PMCs appears normal, diakinesis shows intertwining between the non-homologous chromosomes (Fig. 6E). At metaphase I, multivalent links involving three or more chromosomes form alongside univalents (Fig. 6F). Bridges between chromosome groups and chromosome fragmentation are observed throughout anaphase I (Figs. 6G and 6H), metaphase II (Fig. 6I), anaphase II (Fig. 6J), and telophase II (Fig. 6K). Telophase II cells also exhibit unequal chromosome numbers (Figs. 6G, 6I, 6J, 6K).

We developed the atm×top1α1 double mutant and examined its chromosomal behavior. Compared tothe single-atm mutant, we observed incomplete synapsis of homologous chromosomes at zygotene and pachytene (Figs. 7A–7C and 7D, 7E, respectively). Additionally, some DNA fragments fail to segregate to cell poles at anaphase I (Fig. 7F). These defects lead to the production of multiple polyads (five unbalanced microspores) at telophase II in the double mutant (Figs. 7G–7I).