Introduction

Micronuclei are cell structures that originate from chemical, physical or biological mutagens. Micronuclei are small extranuclear structures caused by clastogenic and/or aneugenic effects (Fig. 1). Clastogenic agents promote lesions in the DNA double strand and the inhibition of the DNA damage repair system, resulting in structural chromosome alterations and/or chromosome bridges1. As a result of the clastogenic effect, micronuclei are generated from acentric chromosome fragments2. The aneugenic effect is characterized by the loss of a whole chromosome, which does not attach to the spindle fibers during mitotic/meiotic metaphases causing anaphase segregation errors3. Defects in microtubule-kinetochore connections promote the aneugenicity as a consequence of cytosine hypomethylation in centromere DNA4,5, incorrect mitotic spindle assembly6, errors in cell cycle check points and/or abnormal centrosome duplication7. The chromosome fragments or lagging chromosomes are delimited by the nuclear envelope separately of the main nucleus at the telophase end, generating the micronuclei1,2.

Fig. 1
figure 1

Micronuclei formation pathways. (a) The clastogenic effect promotes chromosome fragmentation. (b) The aneugenic effect impedes a whole chromosome to attach to the spindle fibers. The micronuclei result from these (a) fragments and/or (b) lagging chromosomes at the telophase end. Figure 1 was created using the Photoshop CC version 2019.

Full size image

Micronuclei can have different fates according to their formation pathway and genome composition, as well as to cell type and species8. The possible fates for these structures include degradation of the micronuclei, reincorporation into the main nucleus, extrusion from the cell, persistence in the cytoplasm, premature chromosome condensation/chromothripsis and elimination by apoptosis8. Due to micronuclei formation pathway and their possible persistence in the cytoplasm, micronuclei frequency is widely used to evaluate the genotoxic effects of different agents in animals9 and plants10.

Considering that micronuclei can be originated from chromosome fragments or whole chromosomes, these structures can have different DNA content11. Since micronuclei can persist in a cell or be lost, it is relevant to know how much DNA the micronuclei possess. Hence, the DNA content measurement of the micronuclei gives additional information on the extent of the damage promoted by genotoxic agents. DNA content measurement in plants12 and animals13 is usually accomplished using flow cytometry. The DNA content of micronuclei from mouse fibroblasts was measured using flow cytometry after treatment with tear gas chlorobenzylidene malonitrile, vinblastine and γ radiation11. Therefore, the large DNA content variation among the micronuclei was evidenced by biparametric histograms (Side-scattered light—SSC x fluorescence intensity). In addition, the DNA content variation of the micronuclei was compared to the DNA content of the chromosomes, which was measured through flow karyotyping. Based on DNA content values (micronuclei and mouse chromosomes), Nüsse et al.11 appointed the micronuclei origin from chromosome fragments or whole chromosomes.

In plant mutagenesis, the micronuclei DNA content has not been measured. The DNA content measurement in plant micronuclei using flow cytometry presents some obstacles, such as the separation of the micronuclei from the cellular wall debris and organelles, like mitochondria and plastids. In addition, flow cytometry requires a large number of particles with the same C value to obtain histograms that allow the recognition of the micronuclei peak14. As micronuclei can present different DNA content values as a consequence of their origin, the measurement of these data is a challenge. In this scenario, image cytometry is an alternative approach for micronuclei DNA measurement. Image cytometry is applied in studies concerning cancer pathology from human cells, allowing the identification of malignant cells from the euploidy/aneuploidy level measurement15. Image cytometry has also been applied to measure the nuclear DNA content in animals16 and plants, as well as chromosome DNA content in plants17,18. Image cytometry has advantages over flow cytometry since it allows the visual recognition of the nuclei19 and the analysis is possible with a small number of nuclei20. In that regard, the measurement of micronuclei DNA content by image cytometry may provide concrete evidence on how much of the nuclear genome can be lost during micronuclei formation.

Micronuclei diversity has also been evidenced by occurrence of DNA sequences from particular chromosome and chromosome portions. Human lymphocytes treated with physical (γ ray radiation21) and chemical mutagens (mitomycin C and diethylstilbestrol22) generated micronuclei composed of sequences originating from different human chromosomes. Similarly, the micronuclei from root meristems of Hordeum vulgare L.23 and Brachypodium distachyon (L.) P. Beauv.24, which were treated with chemical (N-nitroso-N-methylurea and maleic hydrazide) and physical mutagens (γ and X ray radiation), presented sequences of the centromere, telomere, 5S and 25S rDNAs.

Plant genomes present repetitive DNA sequences (repeatome), like tandem repeats and mobile elements25. Moreover, repeatome sequences are considered hotspots for structural chromosome rearrangements. Heterochromatin portions exhibiting repetitive DNA sequences26, mobile elements27 or 180-bp knob sequences28 were identified as regions prone to suffer chromosome rearrangements, especially deletions. Tandem repeated sequences (rDNAs, centromere and telomere) were identified in micronuclei of H. vulgare23 and B. distachyon24, suggesting that repeatome sequences can be involved in micronuclei composition. Furthermore, plant genomes are continuously exposed to abiotic and biotic stress due to environmental changes, which can promote genetic or epigenetic diversity29. Thus, the detection of repeatome sequences allow us to understand the structure and organization of the micronuclei genome30.

In addition to the micronuclei composition, understanding their origin can also help to elucidate questions concerning the structure and organization. Probes generated from microdissected micronuclei can be employed to reveal the micronuclei origin31. A spontaneous micronucleus in human lymphocytes was microdissected and its DNA was amplified by Random Amplified Polymorphic DNA-PCR (RAPD-PCR). From fluorescent in situ hybridization (FISH), the DNA probe revealed that the micronucleus originated from chromosome 232. To our knowledge, a study aiming to find out the micronuclei origin has not been conducted in plant mutagenesis until now.

Zea mays L. is a species with a relatively well-characterized genome, and sequencing data available from public domain databases33. Z. mays genome consists of ~ 85% of mobile elements, which have been characterized33 and mapped28,34. Besides, procedures for chromosome DNA content18 and construction of specific probes from microdissection35 and repeatome sequences28 are available for Z. mays.

We aimed to measure the DNA content, and investigate the composition and origin of Zea mays micronuclei resulted from DNA damage induced by methyl methanesulfonate (MMS). For these, we applied image cytometry, FISH from repeatome sequences (5S and 18S rDNAs, 180-bp knob, Grande LTR-retrotransposon,), 6-diamidino-2 phenylindole (DAPI) staining, and microdissection for micronuclei probe construction.

Results

Micronuclei have widely been used as evidence of DNA damage. However, the mutagenic effect is mainly demonstrated by micronuclei presence and/or frequency. Here, we showed that the micronuclei can be more explored, especially the DNA content lost due to the damage. Besides, the DNA content, composition and origin of the micronuclei provided some insights of how the genome was affected by a mutagen.

The hydrolysis using 5 M HCl for 45 min resulted in morphologically preserved nuclei (prophases and telophases) and micronuclei, without cytoplasmic background and stoichiometrically stained by Schiff’s reagent (Fig. 2; Supplementary Fig. 1). In addition, metaphases with lagging chromosomes were also observed in low rate. System calibration was appropriate, since the obtained 4C/2C ratio, coefficient of variation and standard deviation values were adequate. The 4C/2C ratio oscillated from 2.054 to 2.175, given that the ratio required optimum values lie between 1.8 and 2.215. Coefficient of variation ranged from 2.533 to 4.538% for prophases and 2.347–4.861% for telophases. The values of coefficient of variation for image cytometry analyses must be lower than 5%36. Thus, we considered the image cytometry procedure adequate to measure the IOD values of the prophases, telophases and micronuclei. Standard deviations ranged from ± 3.800 to 4.344 for prophases and ± 2.837–4.419 for telophases, showing the relevance of the MMS treated root meristem and of the negative control in the same slide.

Fig. 2
figure 2

Representative (a) micronuclei, (b) telophases and (c) prophases of Z. mays obtained from root meristems (a) treated and (b, c) untreated with MMS. Micronuclei and nuclei were stained from Feulgen reaction and used to measure the IOD and, consequently, the DNA content of each micronucleus. Bar = 10 µm.

Full size image

For each slide, the IOD values of the micronuclei ranged from 18.168 to 23.207; 24.941–39.494; 14.362–34.169; 16.317–38.297; 19.861–43.051; 19.345–45.624; 24.941–41.647; 11.753–42.259; 11.776–42.960; 13.882–68.086. Each IOD value was converted in DNA content, and the descriptive statistical analysis categorized the micronuclei in 6 groups according to DNA content (Fig. 3). The majority of the micronuclei (69%) presented DNA content ranging from 0.5 to 1.5 pg (groups 2 and 3). Whereas, micronuclei with DNA content ranging from 2.0 to 3.0 pg (groups 5 and 6) were less frequent (4%).

Fig. 3
figure 3

Graphic showing the distribution of the micronuclei DNA content values in pg. Based on DNA content, the micronuclei were categorized in groups 1–6 according to their respective intervals. A predominance of micronuclei with 0.5–1.5 pg was observed. Figure 3 was created using Rbio: Biometria no R, version 210, https://biometria.ufv.br/infra-estrutura/.

Full size image

Sequences of the 5S and 18S rDNAs, 180-bp knob and/or Grande LTR-retrotransposon were detected in the micronuclei. The proportion of micronuclei with signals is summarized in Table 1 for each repeatome sequence. 5S and 18S rDNAs sequences exhibited strong signals on nuclei and micronuclei (Fig. 4a–d). Micronuclei without rDNA sequence signals were also observed (Table 1). As expected, 5S rDNA was observed in the distal region of the long arm of chromosome 2 (Fig. 4a). 18S rDNA was detected in the nucleolar organizer region on the short arm of chromosome 6 (Fig. 4c). 180-bp knob sequence exhibited disperse signals on micronuclei with a variation of fluorescence signals (Fig. 5a, Table 1). In metaphases, positive signals were visualized in the long arm of the chromosomes 1, 2, 4, 5 and 8 and in the short arm of the chromosome 5, with heterozygosity in pairs 1 and 5 (Fig. 5b). We found scattered signals of the Grande LTR-retrotransposon sequence with variable distribution in each micronucleus (Fig. 5c). Moreover, strong uniform signals were observed along the length of the ten chromosomes (Fig. 5d). Micronuclei counterstained with 6-diamidino-2 phenylindole (DAPI) exhibited a variation of positive signals (DAPI+), showing that heterochromatic AT rich-regions are involved in micronuclei composition (Fig. 6).

Table 1 Proportion of Z. mays micronuclei exhibiting at least one fluorescence signal for each repeatome sequence and for DAPI fluorochrome.
Full size table
Fig. 4
figure 4

Genomic composition of the Z. mays micronuclei. Presence of 5S rDNA fluorescence signals on (a) nuclei and chromosome 2 long arm, and (b) micronuclei. 18S rDNA fluorescence signals on (c) nuclei and chromosome 6 short arm, and (d) micronuclei. The 5S and 18S rDNAs fluorescence signals on micronuclei and chromosomes are indicated by the arrow. Bars = 10 µm.

Full size image
Fig. 5
figure 5

Genomic composition of the Z. mays micronuclei. 180-bp knob signals dispersed on (a) nuclei/micronuclei, showing different spots and fluorescence intensity between the micronuclei. (b) 180-bp knob sequence was mapped in the long arm of the chromosomes 1 (heterozygosity), 2, 4, 5 and 8, and in the short arm of the chromosome 5 (heterozygosity). Based on this, (a) micronuclei possess sequences of these Z. mays chromosomes. Scattered Grande LTR-retrotransposon signals on (c) nuclei/micronuclei and on (d) all Z. mays chromosomes, indicating that the micronuclei may be composed of fragments of the ten Z. mays chromosomes. Bars = 10 µm.

Full size image
Fig. 6
figure 6

Genomic composition of the Z. mays micronuclei. Presence of DAPI + signals (arrow) in Z. mays micronuclei evidencing the occurrence AT rich-portions. The number and intensity of the DAPI + signals varied in each micronucleus. Bars = 10 µm.

Full size image

From micromanipulation procedure, we obtained micronuclei presenting DNA fragments ranging from 100 to 1000 bp, predominating 200–400 bp for four microdissected micronuclei, and 100–2000 bp for one micronucleus (Supplementary Fig. 2). Based on this, we selected two micronuclei, which showed the distinct profile in 1.5% agarose gel: micronucleus 1 and micronucleus 2 (Supplementary Figs. 2 and 3). Corroborating the previous data (micronuclei DNA content and DNA sequence composition), the origin of these structures involved different chromosome fragments. Strong hybridization signals were visualized in all chromosomes (Fig. 7). The micronucleus 1 probe pool provided strong signals along the length of all chromosomes, with a lower accumulated profile in chromosome 6 and the short arm of chromosomes 3, 4 and 9 (Fig. 7a). The micronucleus 2 probe pool showed intense spread signals along the length of all chromosomes except on centromere, knob and secondary constriction (Fig. 7b).

Fig. 7
figure 7

Z. mays chromosomes hybridized with micronucleus probe pool labelled with (a) ChromaTide-488–5-dUTP (green) for micronucleus 1 and (b) Tetramethyl-rhodamine 5-dUTP (red) for micronucleus 2 in metaphase. (a) Micronucleus 1 probe pool resulted in scattered hybridization signals on chromosomes, except on chromosome 6 and on short arm of chromosomes 3, 4 and 9. (b) Micronucleus 2 probe pool showed scattered hybridization signals on chromosomes, except on centromere, knob and secondary constriction. Bar = 10 µm.

Full size image

Discussion

Zea mays micronuclei varied in DNA content, composition based on different fluorescence signals of the repetitive DNA sequences, and also origin, revealing the diversity in micronuclei formation. The procedures employed here enabled the ascertainment of the DNA content, as well as expanded the knowledge about micronuclei composition and contributed to some insights about micronuclei origin. Besides, other species, as H. vulgare23 and B. distachyon24, also presented micronuclei with different compositions in the presence of mutagens inducing micronuclei formation.

Additionally, the micronuclei variability observed here results from MMS’s mechanism of action. MMS is a clastogenic and alkylating agent that promotes methylations in DNA nitrogenous bases. This mutagen acts by generating over 80% of its methyl adducts at the N7-guanine, 10% of this alkylation change at the N3-adenine and 0.3% at the O6-guanine37. These purine bases occur along the Z. mays genome, which contains 52.8% AT and 47.2% GC38. Therefore, the micronuclei with different DNA content, repeatome compositions and/or origins are consequence of the MMS damages on several genomic portions of the Z. mays genome.

Based on image cytometry data, the micronuclei are distinct in relation to DNA content. Considering the nuclear 1C value of Z. mays ‘AL Bandeirante’ (1C = 3.05 pg equivalent to 1C = 2.983 × 109 bp18, since 1 pg = 0.978 × 109 bp17), ~ 16.76 to 50.28% of this reference value was lost due to micronuclei formation. Regarding the 1C value of each chromosome of Z. mays ‘AL Bandeirante’18, the micronuclei presented DNA content values equivalent to different chromosome fragments. Moreover, the micronuclei presenting DNA content values between 2.0 and 3.0 pg (Fig. 3, Supplementary Fig. 4) were generated from genomic fragments of different chromosomes (Figs. 4, 5, 6 and 7). The MMS overdose applied here possibly potentialized the number of chromosome breaks, which were accumulated and delimited by the membrane of a single one cell, generating these micronuclei with a high DNA content value. Furthermore, the MMS promotes a delay of the replication fork progression due to methylation at the N3-adenine39,40, leading to a slower S phase progression and an interference in the cell cycle. Hence, probably the majority of the micronuclei evaluated here was observed in cells with 2C value.

5S and 18S rDNA signals were detected in 22.5% and 24.7%, respectively, of the micronuclei, and more than 97% showed signals of the 180-bp knob and Grande LTR-retrotransposon sequences (Table 1). However, the number and/or intensity of the fluorescence signals varied among the micronuclei for all repeatome sequences. Until this study, sequences of the centromere, telomere and rDNA were used to analyze the composition of plant micronuclei23,24. Our results demonstrated that the micronuclei also contain other DNA sequences, such as the 180-bp knob and Grande LTR-retrotransposon sequences. From micronuclei composition and chromosome mapping of the 180-bp knob sequence and DAPI+ banding, we concluded that micronuclei contain fragments of the chromosomes 1, 2, 4, 5 and 8. Moreover, the hybridization profile of the Grande LTR-retrotransposon sequence confirmed the micronuclei composition and revealed that they also contain genomic sequences from other Z. mays chromosomes, as this retrotransposon sequence is distributed across all ten chromosomes. The rDNA of Z. mays is rich in GC content (> 70%), as indicated by its 5S (DQ351339.1, https://www.ncbi.nlm.nih.gov/nuccore/DQ351339.1) and 18S (U42796.1, https://www.ncbi.nlm.nih.gov/nuccore/U42796.1) sequences. 180-bp knob is rich in AT41,42 and the Grande LTR-retrotransposon presents 52% GC43,44. Considering that adenine and, mainly, guanine are targeted by MMS37, these sequences can be indicated as hotspots for MMS-induced DNA damage and consequently micronuclei formation in Z. mays, mostly the Grande LTR-retrotransposon, 5S and 18S rDNAs.

The conclusions reached so far can be supported with results obtained from probe-pools of the microdissected micronuclei. The two-probe pools exhibited hybridization signals across all chromosomes, indicating that these micronuclei were originated from fragments containing sequences common to all Z. mays chromosomes. Based on these findings, the results indicated that micronucleus 1 and micronucleus 2 contain Grande LTR-retrotransposon sequences, as this sequence is evenly distributed across all ten chromosomes of Z. mays. The preference of MMS for guanine37 may have contributed to these results since the Grande LTR-retrotransposon sequence is rich in GC portions43,44. LTR-retrotransposons compose ~ 75% of the Z. mays ‘B73’ reference genome33. Besides the Grande LTR-retrotransposon of the Gypsy superfamily, there are other families of these mobile elements distributed throughout the Z. mays chromosomes, such as Huck, Cinful, Tekay/Prem-1 (Gypsy superfamily) Prem-2/Ji and Opie (Copia superfamily)34. In light of this, the micronuclei may have originated from fragments containing sequences from these other retrotransposons as well.

The different analysed micronuclei presented variable number of specific signals for the 180-bp knob sequence and AT-rich regions (DAPI+). Considering that MMS introduces a methyl group to N3-adenine45, the presence of these signals further reflects the effect of the applied mutagen. Knobs are heterochromatic regions composed of tandem repeat sequences integrated with LTR-retrotransposons34. Therefore, variations in heterochromatin distribution among micronuclei may be associated with the presence of these repetitive DNA sequences in their composition.

Micronucleus 1 and micronucleus 2 originated from different chromosome fragments, as evidenced by the distinct pattern of hybridization signals from each micronucleus probe pool. The micronucleus 1 probe pool showed fluorescence signals frequently distributed across all ten chromosomes, while the micronucleus 2 probe pool did not exhibit hybridization signals in the centromere, knob and secondary constriction. Furthermore, the different origins of micronuclei are further supported by the distinct profiles observed in the products of PCR using degenerated primers (DOP-PCR) for each micronucleus.

Conclusions

MMS induced Z. mays micronuclei with a remarkable genomic diversity about DNA content, repeatome composition (5S and 18S rDNAs, 180-bp knob, Grande LTR-retrotransposon, DAPI +) and origin. Therefore, the micronuclei diversity was demonstrated by image cytometry and FISH using probes generated from repeatome sequences and microdissected micronuclei (Fig. 8). Fragments consisting of 180-bp knob, Grande LTR-retrotransposon and DAPI + were highlighted as frequently involved in micronuclei formation. In addition, the results are correlated to the MMS’s mechanism of action, demonstrating the extent of the genotoxic damage promoted by this mutagen in the Z. mays genome. Our study evidences that the micronuclei should be further explored in research on DNA damage promoted by mutagens. Evaluating micronuclei provides valuable insights into the extent of genomic damage, revealing both the proportion of the nuclear genome lost and the specific DNA sequences affected as a result of mutagen-induced disturbance.

Fig. 8
figure 8

Overview about the techniques and Z. mays genomic sequences used to demonstrate the DNA content, composition and origin of the micronuclei. Figure 8 was created using the Photoshop CC version 2019.

Full size image

Materials and methods

Plant material and micronuclei induction

Commercial seeds of Z. mays ‘AL Bandeirante’ (Sementes Caiçara®, allotment seed number 91, category S2, harvest 2016, germination rate of 94.0%, 99.5% purity) were germinated in Petri dishes on filter paper moistened with dH2O at 30 °C. Methyl methanesulfonate (MMS, Sigma-Aldrich®, CAS No 66–27-3) is a chemical mutagen able to cause DNA damage like strand breaks and adducts, which are expressed as chromosomal rearrangements46. Based on the MMS concentration usually applied as positive control in mutagenesis studies, mostly with Allium cepa L. (4 × 10–4 M47,48,49), and the thickness of the Z. mays root meristems, we used an overdose of MMS to induce a high frequency of micronuclei. For this, roots with 1 cm length were transferred to Petri dishes containing 2 × 10–3 M MMS and incubated for 24 h at 30 °C (~ 20 seeds/Petri dish). Noteworthy, we also tested higher doses than the one used here, which proved lethal to the biological material. Roots germinated only in dH2O were used as negative control. The root tips were washed in dH2O, fixed three times in 3:1 methanol: acetic acid (Merck®) with changes after each 10 min and stored at − 20 °C.

Micronuclei DNA content by image cytometry

After 24 h in 3:1 methanol:acetic acid, the roots were transferred to 95% ethanol and stored at − 20 °C for 24 h. The roots were washed three times in dH2O, macerated in enzymatic pool (4% cellulase Sigma®, 0.4% hemicellulase Sigma®, 1% macerozyme Onozuka R10 Yakult®, 100% pectinase Sigma®) diluted in dH2O in the proportion 1:8 (enzyme pool: dH2O) for 2 h at 36 °C, washed three times with dH2O, fixed three times in 95% ethanol with changes after each 10 min, and stored at − 20 °C for 24 h18. The slides were prepared using the squashing technique. On the same slide, MMS-treated root meristems and dH2O-treated root meristems (negative control) were squashed on opposite sides to ensure both receive exactly the same Feulgen reaction conditions. The slides were frozen at − 80 °C for 10 min, the coverslips were removed and immediately incubated for 12 h in 17:5:1 methanol:formaldehyde:acetic acid (Merck®) at 25 °C. Next, the slides were washed in dH2O, air-dried and hydrolyzed in 5 M HCl (Merck®) for 40–50 min at 25 °C, washed in dH2O, air-dried, and stained with Schiff’s reagent50 for 12 h at 4 °C. Subsequently, the slides were washed in 0.5% SO2− water (Merck®) for 3 min and two times in dH2O for 3 min51.

The setup and calibration of the microscope and image analysis system were performed to measure the optical density (OD) and the integrated optical density (IOD) values50,51. The software algorithm automatically calculated the IOD values multiplying the nuclear area of the nuclei or the micronuclei (µm2) by the mean OD. Ten slides were prepared, and 20 frames were randomly captured for each of them, being 10 for the MMS treatment and 10 for the negative control, resulting in 200 frames in total. The frames were captured using a 12-bit CCD digital video camera (Olympus® DP71) coupled to a trinocular photomicroscope Olympus™ BX-60, equipped with stabilized light source, UPlanFI objective with magnification × 100, 1.4 numeric aperture, aplanat achromat condenser with 1.4 aperture, and neutral density filter. Using the Image Pro-Plus® 6.1 analysis system (Media Cybernetics®), the micronuclei, early prophases and late telophases were digitally segmented. Based on recommendations for image cytometry52, the DNA content was measured from at least ten micronuclei, early prophases and late telophases for each slide. The early prophases and late telophases were selected because they can be visually identified and, more importantly, used to define their ploidy levels.

We calculated the IOD for each prophase and each telophase, and then we calculated the mean IOD for each slide. From these values, we calculated the 4C/2C ratio, coefficient of variation and standard deviation for each slide15,36,52. For each slide, the IOD was also calculated for each micronucleus. Using the formula, we calculated the DNA content in pg from each micronucleus (MN) considering the mean IOD from telophases, the IOD from each micronucleus and the nuclear DNA content of Z. mays ‘AL Bandeirante’ 2C = 6.10 pg18:

$$MN pg= \frac{IOD\,MN \times 6,10 pg}{mean\,IOD\,telophase}.$$

The DNA content values were submitted to a descriptive statistical analysis using Rbio software53 (Rbio: Biometria no R, version 210, https://biometria.ufv.br/infra-estrutura/) to group micronuclei.

Probe construction

Micronuclei composition was investigated using probes from repeatome sequences: 5S and 18S rDNAs, 180-bp knob and Grande LTR-retrotransposon. These sequences were used due to their occurrence in the Z. mays genome. The probes were constructed via PCR using the primers listed in the Supplementary Table 1.

For micronuclei origin, we separately microdissected five micronuclei and generated a specific probe pool. For this, we prepared five slides from macerated MMS-treated root meristems. Each meristem was dissociated under a slide, and the slides were immediately immersed in dH2O for 1 min to keep the micronuclei hydrated and to allow the material removal without chromatin fragment loss. The microdissections were carried out using an Eppendorf TransferMan® micromanipulator coupled to an inverted phase contrast microscope IX70 (Olympus™) in a room at a relative humidity of ~ 70%. The slides were visualized with an objective LUCPlanFLN UIS 2 60X/0.70 Ph2 and five micronuclei were microdissected using sterile Femtotips (Eppendorf®) glass microneedles. Each micronucleus was individually transferred to sterile microtubes of 0.2 mL containing 0.1 mL of sterile collection solution (10 mM Tris–HCl pH 7.5 + 10 mM NaCl + 1 mM EDTA + 0.1% SDS + 0.1% Triton X-100 + 30% Glycerol + 1.44 mg mL−1 Proteinase-K, Sigma®). Immediately, the micronuclei were deproteinized for 2 h at 60 °C, followed by enzyme inactivation at 80 °C for 20 min54. All steps were conducted in UV-irradiated biohazard flow chambers, and the pipetting procedures were carried out with tips containing sterile filters in order to avoid contamination35.

After the Proteinase-K treatment, the genomic DNA of each micronucleus was amplified and labeled by PCR using a DOP primer 5′–CCGACTCGAGNNNNNNATGTGG–3′55. As recommended35, this DOP primer was used since we do not know the genomic composition of each microdissected micronucleus. DNA amplification was accomplished in two rounds35,54. The first amplification reaction mix consisted of the deproteinized micronucleus DNA, 4 µM DOP primer, 200 µM of each dNTP (Promega®), 1X Thermo Sequenase DNA Polymerase (GE®) reaction buffer and 20X Thermo Sequenase DNA Polymerase (GE®). The PCR conditions were initial denaturation at 95 °C for 3 min; 10 cycles of 92 °C for 1 min; 35 °C for 2 min; ramp of 0.1 °C/s until 72 °C; 72 °C for 1:30 min; 30 cycles of 92 °C for 1 min; 56 °C for 2 min; 72 °C for 2 min and final extension at 72 °C for 5 min. The second amplification reaction mix consisted of 200 ng of the amplified micronucleus DNA, 4 µM DOP primer, 200 µM of each dNTP (Promega®), 1X Thermo Sequenase DNA Polymerase (GE®) reaction buffer and 20X Thermo Sequenase DNA Polymerase (GE®). The PCR conditions were initial denaturation at 96 °C for 3 min; 30 cycles of 91 °C for 1 min; 56 °C for 1 min; ramp of 0.1 °C/s until 72 °C; 72 °C for 2 min and final extension at 72 °C for 5 min. The amplification products were visualized by electrophoresis in 1.5% agarose gel. Based on differences in the sizes of the amplified fragments (Supplementary Fig. 2), we chose two genomic micronuclei DNA, which were denominated as micronucleus 1 and micronucleus 2 (Supplementary Figs. 2 and 3).

The labeling reaction for the probe pool construction consisted of 200 ng of the amplified DNA, 4 µM DOP, 200 µM each of dATP, dCTP and dGTP, 100 µM dTTP, 1X Thermo Sequenase DNA Polymerase (GE®) reaction buffer, 20X Thermo Sequenase DNA Polymerase (GE®) and 50 µM fluorochromes ChromaTide® Alexa Fluor® 488-5-dUTP (Life Technologies®) for micronucleus 1 and Tetramethyl-rhodamine 5-dUTP (Roche®) for micronucleus 2. The amplification conditions for the labeling reaction were identical to the second round described above. All amplification and labeling mixes were prepared in UV-irradiated flow chambers, using sterile pipettes and tips with filter, to avoid DNA contamination.

Composition and origin of the micronuclei

We verified the occurrence of 5S and 18S rDNAs, 180-bp knob and Grande LTR-retrotransposon sequences in nuclei and micronuclei. For this, enzymatically macerated MMS-treated root meristems (section “Micronuclei DNA content by image cytometry”) were fixed in 3:1 methanol:acetic acid. Slides were prepared from the cellular dissociation of these roots and air-dried28.

In addition, slides with metaphases were used to map the 5S and 18S rDNAs, 180-bp knob and Grande LTR-retrotransposon sequences and to verify the origin of the microdissected micronuclei. Seeds were germinated in dH2O, and the 1 cm roots were incubated at 30 °C for 18 h in solution of 1.75 mM hydroxyurea (Sigma®) and 0.20 g L−1 MS salts (Sigma®) for cell cycle synchronization. The roots were washed for 1 h in dH2O at 30 °C with changes every 15 min for cell cycle recovery. For metaphase arresting, the roots were treated with 3 µM amiprophos-methyl (Sigma®) in 0.3% dimethyl sulfoxide (Sigma®) for 4 h at 30 °C. The roots were washed in dH2O and fixed three times in 3:1 methanol:acetic acid with changes after each 10 min and stored at − 20 °C. The roots were washed three times in dH2O, submitted to enzymatic maceration (section “Micronuclear DNA content by image cytometry”), washed in dH2O, fixed in 3:1 methanol: acetic acid with three changes after 10 min, and stored at − 20 °C. Slides were prepared by cellular dissociation and air-drying techniques. Slides with morphologically preserved metaphases, nuclei and micronuclei were aged for at least 5 days28.

The slides were washed in 1X phosphate buffered saline (PBS) buffer for 5 min, fixed with 4% formalin for 15 min, washed again in 1X PBS for 5 min, and dehydrated in cold ethanol series (70%, 85% and 100%) for 5 min each. For the 5S and 18S rDNA probes, the hybridization mix consisted of 50% formamide (Sigma®) + 2X saline-sodium citrate (SSC, Sigma®) and 100 ng of the probe56. For the 180-bp knob and Grande LTR-retrotransposon probes, the hybridization mix consisted of 50% formamide (Sigma®) + 2X SSC (Sigma®), 35 µg competitor DNA (Herring Sperm DNA, Promega®) and 100 ng of the probe28. For the micronuclei probe pool, the hybridization mix consisted of 50% formamide (Sigma®) + 2X SSC (Sigma®), 20 µg competitor DNA (Herring Sperm DNA, Promega®), 200 ng extra competitor DNA (amplified DNA from micronucleus 1 for micronucleus 2 probe pool and conversely) and 200 ng of the probe pool. The mixes were denatured at 85 °C for 5 min, followed by immediate transfer to ice for at least 5 min. 35 µL of the mix were placed on the slides, which were covered by a plastic coverslip HybriSlip (Sigma®) and sealed with Rubber Cement Elmer’s®. In a ThermoBrite system (ThermoFisher®), nuclei and micronuclei were denatured at 68 °C for 3 min and metaphases at 70 °C for 3 min, followed by hybridization at 37 °C for 24 h. After, the coverslips were removed and stringency washes were performed in three solutions of 50% formamide/2X SSC and one of 2X SSC, for 5 min each at 45 °C, resulting in a stringency of ~ 82%. The slides were counterstained with 40% glycerol/PBS + DAPI, covered with 24 × 40 mm glass coverslip and sealed with nail polish57. The micronuclei images obtained from DAPI counterstaining were also used to investigate the presence of AT rich-portions in the micronuclei composition.

For each probe, at least 15 metaphases and 20 nuclei with micronuclei were randomly captured from ten slides with a photomicroscope Olympus™ BX60 equipped with an immersion objective 100×/A.N. 1.4, a 12-bits CCD digital video camera (Olympus™ DP71), WB (micronucleus 1 probe pool—green florescence), WG (micronucleus 2 probe pool, 180-bp knob, Grande LTR-retrotransposon, 5S and 18S rDNAs probes—red florescence) and WU (DAPI staining) filters, and a computer with a digitizer plate. The images were captured using the exposure time of 1/1.8 s for green and red florescence and 1/60 s for DAPI staining, processed by Image ProPlus 6.1 software (Media Cybernetics®) and edited with the same brightness and contrast in Adobe Photoshop CC.