Repetitive sequences and structural chromosome alterations promote intraspecific variations in Zea mays L. karyotype

LTR-retrotransposons, knobs and structural chromosome alterations contribute to shape the structure and organization of the Zea mays karyotype. Our initial nuclear DNA content data of Z. mays accessions revealed an intraspecific variation (2 C = 2.00 pg to 2 C = 6.10 pg), suggesting differences in their karyotypes. We aimed to compare the karyotypes of three Z. mays accessions in search of the differences and similarities among them. Karyotype divergences were demonstrated among the accessions, despite their common chromosome number (2n = 20) and ancestral origin. Cytogenomic analyses showed that repetitive sequences and structural chromosome alterations play a significant role in promoting intraspecific nuclear DNA content variation. In addition, heterozygous terminal deletion in chromosome 3 was pointed out as a cause of lower nuclear 2 C value. Besides this, translocation was also observed in the short arm of chromosome 1. Differently, higher 2 C value was associated with the more abundant distribution of LTR-retrotransposons from the family Grande in the karyotype. Moreover, heteromorphism involving the number and position of the 180-bp knob sequence was found among the accessions. Taken together, we provide insights on the pivotal role played by repetitive sequences and structural chromosome alterations in shaping the karyotype of Z. mays.

The DNA content has been measured for the long and short arms of each chromosome and the satellite of chromosome 6 of Z. mays ' AL Bandeirante' 15 . The DNA content of chromosome 9 (2 C = 0.56 pg) was higher than that of chromosome 8 (2 C = 0.53 pg) 15 , a fact that can be related to accumulation of repetitive sequences in the knobs 16 . Hence, intraspecific variation in nuclear or chromosomal DNA content hints at karyotype differences, emphasizing the need to understand the causes of these variations in distinct Z. mays accessions.
In addition to TEs, knobs are also responsible for the intraspecific variation in Z. mays nuclear genome size 17,18 . Knobs are heterochromatic regions identified in pachytene and mitotic prometaphase and metaphase chromosomes by means of differential staining techniques 19 . They are composed of two tandem repeat sequences, of 180 base pairs (bp) 20 and 350 bp (TR-1), besides harboring several LTR-retrotransposons 17,21 . The positions, number and size of the knobs are variable among both the accessions and the chromosomes of the same karyotype 9,22 . In some cases, the knobs might serve as chromosomal markers that provide physical evidence of crossing-over events between non-homologous chromosomes 23 . Despite the recognized role of TEs and knobs on the dynamism of Z. mays genome size variation, little is known about the important of these sequences on plant fitness. However, Bilinski et al. 18 have found evidences that this variation may indeed be adaptive and that heterochromatic knob sequences are likely under the effect of natural selection. Therefore, considering that knob heteromorphism correlates with nuclear genome size, it is fundamental to map these portions in Z. mays chromosomes in order to verify their involvement in DNA content divergence.
In addition to LTR-retrotransposons and knobs, the chromosome structure and morphology can also be altered by structural rearrangements: duplication, deletion, translocations and/or inversions 24 . These rearrangements have been revealed in the genera Solanum 25 , Brachypodium 26 and Z. mays 27 by comparative cytogenetics via chromosome painting, thus assisting the elucidation of their evolutionary histories. During the process of double-strand break repair, several rearrangements may occur as a result of illegitimate recombination or through recombination of homologous ectopic sequences 28 . Thus, repetitive sequences such as TEs can provide a template for repairing the double-strand breaks. For this reason, heterochromatic regions rich in similar repetitive sequences are considered hotspots for double-strand breaks 29,30 .
Beyond the karyotype diversity and dynamism of Z. mays ssp. mays, this taxon also occupies a wide range of habitats and presents a diversity of morphological traits 1,31 , being a crop with several agricultural varieties specific for different uses 31 . For example, the popcorn is characterized by small, hard kernels that explode when heated, forming large flakes (popping expansion), the major feature that separates popcorn from other types of maize 31,32 . Sturtevant 33 considered popcorn as a distinct species, Zea everta, which was posteriorly reduced to a subspecies 34 and then considered as a mutant of flint maize 35 . However, archeological evidence and the quantitative trait of popping ability rendered improbable the hypothesis of a mutant origin from flint maize 32 . Currently, taxonomists consider that popcorn belongs to the taxon Z. mays ssp. mays (https://www.itis.gov/about_itis.html; http://www.plantsoftheworldonline.org). The origin and evolutionary relationship of popcorn with other types of maize remains unknown 31 . Therefore, the genomic in situ hybridization (GISH) and the comparative chromosome painting via chromosome-specific probes might to discriminate the homologous chromosome regions, contributing to understand the evolutionary relationships of popcorn.
Considering the remarkable karyotype dynamism, intraspecific variation in nuclear genome size and chromosomal DNA 15,36 within Z. mays, the aim of this study was to perform a comparative analysis of the karyotypes of different Zea accessions, seeking to identify if the nuclear genome size variation among them is promoted by differential amounts of repetitive sequences and/or by structural chromosomal rearrangements.
Given these nuclear genome size differences, we explored the karyotypes in order to understand the causes of these divergences among Zea accessions. For this, a metaphasic index of 60% was obtained from the cell cycle arrest, achieved by treatment involving hydroxyurea followed by amiprophos-methyl. Besides, to ensure morphologically preserved chromosomes with well-defined telomeres and primary constrictions, enzymatic maceration of root meristems and air-drying technique were used for slide preparation.
Once the differences in nuclear genome size were verified, the genomic homology among Z. mays accessions was confirmed. Even with GISH stringency at 85-90%, all chromosomes of at least ten 'Milho Pipoca Americano RS 20' and ' AL Bandeirante' metaphases were fully hybridized by the genomic probe of Z. diploperennis. The same result was observed for the genomic probe of 'Milho Pipoca Americano RS 20' applied to the ' AL Bandeirante' karyotype ( Fig. 1). From hybridization of our previously constructed probe of Z. mays ' AL Bandeirante' chromosome 1, the homology between Z. mays accessions was also evidenced by specific painting of the chromosome 1 of 'Milho Pipoca Americano RS 20' (Supplementary Fig. 2). Therefore, these results provide substantial evidence that popcorn belongs to the subspecies mays.
Nevertheless, karyotype differences were demonstrated among the Z. mays ssp. mays accessions. Structural chromosome changes were identified in all karyotypes of '15-1149-1' , which were specifically stained by Feulgen reaction, DAPI and labeled by the 180-bp probe. Two alterations were identified ( Supplementary Fig. 3): terminal deletion in the long arm of chromosome 3 and translocation in the short arm of chromosome 1 (Fig. 2a). The translocation was identified as one detectable chromosome fragment, but it was not classified as non-reciprocal or reciprocal. Differently, no structural chromosomal aberration was observed in 'Milho Pipoca Americano RS 20' and ' AL Bandeirante' karyotypes ( Fig. 2b,c).
Apart from these structural chromosome aberrations, karyotype variations were also found regarding the number and position of the 180-bp knob sequence, including heteromorphism within the same karyotype between the chromosome pair (Fig. 3). The 180-bp sequence was mapped in nine different chromosome portions (1 L, 2 L, 3 L, 4 S, 4 L, 5 S, 5 L, 6 L or 8 L) in the karyotypes of Z. mays '15-1149-1' and ' AL Bandeirante' . Z. mays '15-1149-1' exhibited positive signals in the chromosome portions 3 L, 4 S, 5 L, 6 L and 8 L (Fig. 3a,b), whereas ' AL Bandeirante' displayed signals in 1 L, 2 L, 4 L, 5 S, 5 L and 8 L, also presenting heterozygosity for the presence/ absence of signals in the chromosome pairs 1 and 5 (Fig. 3c,d). The karyotypes of the analyzed Z. mays accessions also differed in relation to Grande LTR-retrotransposon mapping. Uniform hybridization signals from this probe were obtained in ten metaphases of 'Milho Pipoca Americano RS 20' (Fig. 4a). The ' AL Bandeirante' karyotype exhibited stronger hybridization signals throughout the chromosome length in 15 metaphases (Fig. 4b), indicating that ' AL Bandeirante' possesses more copies of this LTR-retrotransposon.  In 'Milho Pipoca Americano RS 20' and ' AL Bandeirante' , no structural chromosomal aberration was observed. The DAPIbanding pattern in Z. mays chromosomes was promoted by the preferential binding of this fluorochrome to A-T rich sequences, allowing to evidence the knob portions in a cyan blue color. Bar = 10 µm. Images were digitized using the Image-Pro Plus software version 6.1 (https://www.mediacy.com/imageproplus).

Discussion
Comparing the nuclear DNA content of Z. diploperennis (2 C = 5.76 pg), 'Milho Pipoca Americano RS 20' (2 C = 5.55), '15-1149-1' (2 C = 2.00 pg) and ' AL Bandeirante' (2 C = 6.10 pg), a difference of up to 4.10 pg was found. Considering the chromosomal DNA content of ' AL Bandeirante' , and according to mean values reported by Silva et al. 15 , the 4.10 pg corresponds to approximately five times the chromosome 1 (2 C = 0.80 pg = 4.00 pg) or ten times the chromosome 10 (2 C = 0.38 pg = 3.80 pg). These data reinforce the intraspecific variation in nuclear DNA content observed in Z. mays, which has been reported to range from 2 C = 4.50 pg to 7.11 pg 13,14 . Given this variation, karyotype differences and similarities were sought inside each Z. mays accession studied here.
Based on divergences regarding nuclear genome size, the first step was to verify the evolutionary relationship of the Z. mays spp. mays accessions. Genomic probes of Z. diploperennis provided hybridization signals, from telomere to telomere, in the chromosomes of ' AL Bandeirante' and 'Milho Pipoca Americano RS 20' , confirming the genomic affinity of these accessions to the basal species Z. diploperennis 1 . Hybridization signals were also observed over all Z. mays ssp. mays chromosomes with the genomic probe of Z. diploperennis, but they were weaker than those detected with genomic probes of Z. mays spp. mexicana and Z. mays spp. parviglumis 37 , which are more phylogenetically close to Z. mays spp. mays 1 . Reinforcing the evolutionary relationship, the F 1 hybrids of Z. diploperennis x Z. mays ssp. mays presented regular meiotic chromosome pairing and high pollen viability 38,39 .
The genome homology between ' AL Bandeirante' and 'Milho Pipoca Americano RS 20' was also endorsed by chromosome painting of the chromosome 1. Using the application proposed by Soares et al. 40 , chromosome  www.nature.com/scientificreports www.nature.com/scientificreports/ painting between ' AL Bandeirante' and 'Milho Pipoca Americano RS 20' was informative in comparative karyotype analysis, reflecting the common evolutionary origin of these accessions. This approach has been successfully used for the resolution of phylogenetic questions in Brachypodium genus 26 . The GISH and chromosome painting data obtained here confirmed that popcorn belongs to Z. mays ssp. mays.
After confirming the evolutionary origin, some karyotype divergences were evidenced among the Z. mays accessions, as well as differences between homologue chromosome pairs in the same karyotype, providing cytogenetic data to understand the intraspecific variation in nuclear DNA content. Translocation and terminal deletion were distinguished in '15-1149-1' , which resulted in a morphological change in the chromosomes 1 and 3, respectively. Thus, cytogenetic preparations applying cell dissociation combined with air-drying technique were considered essential for the correct interpretation of these karyotype changes, since this methodology replaces the squashing step that can promote chromosome breakages 41 . The terminal deletion in the long arm of chromosome 3 occurred around the knob, which is considered a hotspot of chromosome structure alterations. The knobs present a complex organization, in which blocks of 180-bp sequence are interrupted by LTR-retrotransposons 21 . Structural chromosomal rearrangements occur within regions composed of repetitive DNA sequences 24,30 . In addition, Lysák and Schubert 42 related that repetitive sequences scattered throughout the genome, especially TEs, are involved in various chromosomal rearrangements, such as deletion and translocation, because ectopic homologous sequences provide a template for recombination during the repair of double-strand breaks, a phenomenon denominated ectopic recombination.
The translocation and terminal deletion found in '15-1149-1' represent one of the causes associated to the relatively lower nuclear DNA content (2 C = 2.00 pg) of this accession in relation to the others. In S. lycopersicum, a difference of 2 C = 0.09 pg between the wild type and the mutant 'BHG 160' was found to be due to a heterozygous terminal deletion in the short arm of the chromosome 1 43 . The translocation is another karyotype aberration that includes a broken chromosome, resulting in a chromatid fragment. However, differently from a deletion or inversion, the chromatid fragment moves and joins the homologue pair or another chromosome 44 . Therefore, the translocation in the short arm of chromosome 1 also evidences that chromosome breakage occurred in the '15-1149-1' karyotype.
These structural chromosome aberrations were highlighted in '15-1149-1' and associated to low nuclear DNA content, yet other karyotype divergences were found from the mapping of the 180-bp sequence and Grande LTR-retrotransposon. Intraspecific variation in nuclear DNA content was also an outcome of the number and heterozygosity of the 180-bp sequence in the karyotype, as well as of the Grande LTR-retrotransposon signals found among the Z. mays accessions. This result shows that repetitive sequences typical of heterochromatin portions, LTR-retrotransposons and 180-bp knobs, are also responsible for genome size variation within Z. mays. The increase in nuclear genome size in this species has also been correlated with 180-bp knob abundance 45 . Furthermore, Bilinski et al. 18 demonstrated that the variations in nuclear genome size are driven by natural selection, causing changes in the abundance of repeat sequences across the genome of Z. mays, as significant reductions in heterochromatic knobs. Knobs, which are constituted by 180-bp and 350-bp sequences, are polymorphic in relation to their number, size and distribution across the ten Z. mays chromosomes 17,46 , affecting 5-20% of the length of the chromosome arm 47 . Besides, the heterozygosity observed in the chromosome portions 1 L and 5 S only in ' AL Bandeirante' is also a karyotype evidence of the differential accumulation of the 180-bp sequence, and consequent change in the chromosomal DNA content among accessions. A heterozygous condition has been appointed as a cause of crossing-over suppression 48 . Although the origin of this polymorphism is still uncertain, it has been presumed that knobs can move around according to the "complex megatransposons" hypothesis 49 . This hypothesis proposes that TR-1 tandem repeat sequences are capable of forming fold-back DNA segments, driving the knobs in the Z. mays genome 49 .
Regarding the distribution of the Grande LTR-retrotransposon, which belongs to the Gypsy superfamily, ' AL Bandeirante' stood out with stronger hybridization signals than 'Milho Pipoca Americano RS 20' . In addition to this mapping, represented by karyograms, the influence of the Grande LTR-retrotransposon sequence on nuclear DNA content among Z. mays accessions was also demonstrated. Grande LTR-retrotransposon distribution produced a uniform hybridization pattern along the extension of metaphasic Z. mays chromosomes, differently from was reported by Mroczek and Dawe 10 and Lamb et al. 9 that found a speckled hybridization pattern along the chromosome extension. Distribution of the Gypsy LTR-retrotransposon in different chromosome regions has also been reported for other species. In Arabidopsis thaliana (L.) Heynh 50 and Asparagus officinalis L. 24 , this LTR is mainly distributed in the centromeres. Differently, in Silene latifolia Poir., the signals for this LTR were observed in subtelomeric heterochromatin regions of the chromosomes 51 .
Many mechanisms have shaped the karyotype organization in plants, such as LTR-retrotransposons 8 . The higher 2 C value and more Grande LTR-retrotransposon signals in ' AL Bandeirante' than in 'Milho Pipoca Americano RS 20' reflect the consequences of the LTR-retrotransposon dynamism -amplification and/or loss. Increase in nuclear genome size is promoted by a de novo DNA sequence of the retrotransposon that is inserted into the genome after an RNA intermediate to be converted into a cDNA molecule by the reverse transcriptase 4,52 . On the other hand, unequal and illegitimate recombination are associated with a high frequency of genomic DNA loss, and may counterbalance the amplification of LTR-retrotransposons 53 , as reported for Arabidopsis 54 and Oryza sativa L. 55 .

conclusions
Cytogenomic analysis showed that the intraspecific nuclear genome size variation in Z. mays spp.  3 , Z. diploperennis is basal in relation to Z. mays spp. mays. Therefore, this species was used to compare the nuclear genome size and to construct genomic probes in order to verify the ancestral relationship and homology among Z. mays spp. mays accessions. nuclear genome size. In order to avoid G 0 /G 1 peak overlapping in flow cytometry histograms due to close nuclear DNA content, the nuclear genome sizes of Z. diploperennis, ' AL Bandeirante' , 'Milho Pipoca Americano RS 20' and '15-1149-1' (samples) were measured using the reference standards S. lycopersicum or Z. mays (2 C = 2.00 pg and 2 C = 5.55 pg, respectively; Praça-Fontes et al. 56 . Leaf fragments from each sample and each internal standard (S. lycopersicum or Z. mays) were co-chopped 57 , and the nuclei were isolated and stained using Otto buffers 58 , following the procedure proposed by Praça-Fontes et al. 56 . The nuclei suspensions were stained with propidium iodide and analyzed in a BD Accuri C6 flow cytometer (Accuri cytometers, Belgium) equipped with a laser source to detect emissions at FL3 (>670 nm). The histograms were analyzed using the BD CSampler software. Four technical replicates were performed for each sample with each standard, analyzing over 10,000 nuclei each time. The mean 2 C nuclear genome size was measured for each Zea sample by dividing the mean channel of the fluorescence peak corresponding to the standard's G 0 /G 1 nuclei by that of each sample.
Due to the intraspecific variation in mean 2 C value among the Z. mays accessions, the karyotypes were characterized with the aim of identifying possible differences and similarities among them. For this, the 180-bp knob sequence and Grande LTR-retrotransposon were mapped via fluorescence in situ hybridization (FISH). In addition, GISH using the genomic DNA of the wild related species Z. diploperennis was performed to confirm the evolutionary origin of popcorn. Owing to constraints in seed availability and low germination rate, '15-1149- Recently, our research group constructed a chromosome-specific probe for the chromosome 1 of Z. mays ssp. mays ' AL Bandeirante' 40 , which was used for chromosome painting in Z. mays. DNA from the chromosome 1, previously amplified by DOP-PCR as reported in Soares et al. 40 , was used in a new labeling reaction. The PCR program and labeling of the amplified fragment were performed as described above. The probe obtained from chromosome 1 was evaluated by electrophoresis in 1.5% agarose gel, showing fragments ranging from 100 to 900 bp. www.nature.com/scientificreports www.nature.com/scientificreports/ Mapping of the 180-bp knob sequence and Grande LTR-retrotransposon. The Grande LTRretrotransposon probe was generated by PCR using the primers F: 5′-TGCGAGGATAAGTCGGCGAAG-3′ and R: 5′-GGTGTTTTTAGGAGTAGGACGGTG-3′ 10 . This family was selected for its wide distribution in the Z. mays genome. The probe of the 180-bp knob sequence was amplified from the primers F: 5′-ATAGCCATGAACGACCATTT-3′ and R: 5′-ACCCCACATATGTTTCCTTG-3′ 14 . The reaction mixture consisted of: 0.5 μM of each primer, 200 ng of genomic DNA, 200 μM of each dNTP (Promega), 1X reaction buffer (Invitrogen), 2 mM MgSO 4 (Invitrogen), and 2 U Platinum Taq DNA Polymerase High Fidelity (Invitrogen). The PCR conditions for the Grande LTR-retrotransposon were as follows: initial denaturation at 95 °C for 5 min; 30 cycles of denaturation at 95 °C for 1 min; annealing at 66 °C for 1 min; extension at 68 °C for 1 min and 30 sec; and final extension at 68 °C for 5 min. For the 180-bp sequence, amplification conditions were the following: initial denaturation at 95 °C for 5 min; 30 cycles of denaturation at 95 °C for 1 min; annealing at 47 °C for 1 min; extension at 68 °C for 1 min and 30 sec; and final extension at 68 °C for 5 min. The labeling reaction consisted of 0.5 μM of each primer, 200 ng of the amplified DNA, 200 μM each of dATP, dCTP and dGTP, 150 μM dTTP, 40 μM Tetramethylrhodamine 5-dUTP (Roche), 1X of the enzyme reaction buffer (Invitrogen), and 2.5 U AccuTaq LA DNA Polymerase. The PCR conditions were the same as described above for each sequence.
in situ hybridization. The procedures were performed as described by Soares et al. 40

and Schwarzacher and
Heslop-Harrison 60 , with modifications. Briefly, the slides were washed in 1X 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. Chromosome denaturation was carried out in 70% formamide/2X saline-sodium citrate (SSC) buffer for 3 min, at 68 °C for 'Milho Pipoca Americano RS' and '15-1149-1' and 70 °C for ' AL Bandeirante' . The difference in temperature is due to over-denaturation when popcorn chromosomes were submitted to 70 °C. Subsequently, the slides were dehydrated in cold ethanol series (70%, 85% and 100%). The hybridization mixture consisted of 50% formamide (Sigma) + 2X SSC (Sigma), 35 μg competitor DNA (Herring Sperm DNA, Promega) and 200 ng of the probe, with denaturation at 85 °C for 5 min followed by immediate transfer to ice. Slides were incubated with 35 μL hybridization mixture, covered by plastic coverslip HybriSlip (Sigma) and sealed with Rubber Cement (Elmer's). The hybridization procedure was conducted in a ThermoBrite system (ThermoFisher) at 37 °C for 24 h. After this period, stringency washes were performed in three solutions of 50% formamide/2X SSC and one of 2X SSC, for 5 min each, at 42 °C for 'Milho Pipoca Americano RS' and '15-1149-1' and 45 °C for ' AL Bandeirante' . Metaphases were counterstained with 40% glycerol/PBS + 6-diamidino-2-phenylindole (DAPI). The same slides used for FISH mapping of the 180-bp knob sequence were also used to evaluate the differential DAPI banding pattern.
The images were captured with a digital video camera 12-bit CCD (Olympus) coupled to a photomicroscope Olympus BX-60 equipped with epifluorescence and immersion objective of 100×, numeric aperture of 1.4. The frame was digitized using the Image-Pro Plus 6.1 software (Media Cybernetics).