Introduction

The development of a karyotype is essential to gain insights into the chromosome organization and evolution of related species. The early 1980s represented the beginning of the molecular cytogenetic era with the advent of FISH (Fluorescent In Situ Hybridization) technique (Langer-Safer et al. 1982). FISH provides a powerful tool for individual chromosome identification by using a wide range of DNA probes, including large-insert genomic DNA clones, such as BACs (Bacterial Artificial Chromosomes) (Jiang and Gill 1994, 2006; reviewed by Jiang 2019). Advances in sequencing technologies, with availability of genome sequences from an increasing number of species, and massive synthesis of oligonucleotides (oligos), have allowed the design of new probes for FISH, such as oligo-painting (Beliveau et al. 2012). These probes can be developed from any species with a sequenced and assembled genome and can be used in karyotype analyses of related species (Jiang 2019).

In plants, chromosome painting with oligo-based probes has been used to determine karyotypes and investigate chromosome rearrangements, meiotic pairing, and recombination in a wide range of species (Han et al. 2015; Qu et al. 2017; Braz et al. 2018, 2021; He et al. 2018, 2020; Hou et al. 2018; Meng et al. 2018; Xin et al. 2018; Albert et al. 2019; do Vale Martins et al. 2019; Šimoníková et al. 2019; Zhao et al. 2019; Bi et al. 2020; Bielski et al. 2020; Li et al. 2020; Liu et al. 2020; Piperidis and D’Hont 2020; Song et al. 2020). Oligos selected from multiple regions from multiple chromosomes can be used to produce a barcode signal pattern for an individual chromosome pair, allowing the identification of a complete set of individual chromosomes in a single FISH experiment (Braz et al. 2018). So far, the oligo-FISH barcode system has been established in a few species, including potato (Braz et al. 2018), maize (Braz et al. 2020a; 2021), rice (Liu et al. 2020), sugarcane (Meng et al. 2020; Piperidis and D’Hont 2020), Ipomoea taxa (Chen et al. 2020), and Triticeae tribe (Li et al. 2020).

Vigna Savi and Phaseolus L. (Leguminosae family, Phaseoloid clade, Phaseolinae subtribe) are phylogenetically related genera, which diverged about 9.7–15 million years ago (Mya) (Li et al. 2013; Garcia et al. 2020). The group includes socioeconomically important species, such as Vigna unguiculata (L.) Walp. (cowpea) and Phaseolus vulgaris L. (common bean), essential crops for global food security and human population nutrition, especially in developing countries (Broughton et al. 2003; Gepts et al. 2008; Freire-Filho et al. 2011; Boukar et al. 2016). Most Phaseolus and Vigna species have 2n = 2x = 22 chromosomes, which are small in size (1–4 μm) and morphological similar varying from meta- to submetacentric chromosomes (Darlington and Wylie 1955; Forni-Martins 1986; Mercado-Ruaro and Delgado-Salinas 1996, 1998, 2000; Venora et al. 1999).

A BAC-FISH-based cytogenetic map was first established for P. vulgaris (Fonsêca et al. 2010). The established BAC markers were used for comparative mapping in Phaseolus lunatus L. (Bonifácio et al. 2012; Almeida and Pedrosa-Harand 2013) and Phaseolus microcarpus Mart. (Fonsêca and Pedrosa-Harand 2013) and revealed a generally high degree of macrosynteny among these species and collinearity breaks due to a few inversions. In turn, in the Leptostachyus group of species (2n = 20), the BAC markers revealed several translocations and a nested chromosome fusion that resulted in descending dysploidy (Fonsêca et al. 2016; Ferraz et al. 2020). BACs from P. vulgaris were also hybridized to V. unguiculata chromosomes and allowed identification of macrosynteny breaks involving a duplication, translocations, and inversions between species (Vasconcelos et al. 2015). Other studies included application of V. unguiculata BACs in genetic and physical map integration (Iwata-Otsubo et al. 2016), besides V. unguiculata and P. vulgaris BACs for BAC-FISH maps for Vigna angularis (Willd.) Ohwi and Ohashi (do Vale Martins et al. 2021) and Vigna aconitifolia (Jacq.) Maréchal (Oliveira et al. 2020). These studies enabled the expansion of the comparative cytogenetic analyses between Vigna and Phaseolus species. Although BAC-FISH is highly informative, it is a laborious process and technically challenging to identify all chromosomes in the same metaphase cells, limiting its application to a small number of species.

The availability of sequenced and assembled genomes for V. unguiculata (1C = 640 Mb; Lonardi et al. 2019) and P. vulgaris (1C = 587 Mb; Schmutz et al. 2014) allows designing oligo-FISH probes for both genomes, as previously demonstrated by oligo-painting on Vigna and Phaseolus species using probes from chromosomes 2 and 3 of P. vulgaris (do Vale Martins et al. 2021). Here, we present a new barcode system for the chromosome identification of both species. We compared the V. unguiculata and P. vulgaris genome sequences and selected V. unguiculata oligo pools from conserved regions in both species so that each chromosome is associated with a unique barcode pattern. This approach will enable comparative FISH mapping and individual chromosome identification of different Vigna and Phaseolus species and potentially other related Phaseolinae members. The accurate chromosome identification will be crucial for future investigation of chromosome evolution and diversification within the subtribe, especially for the non-model species with neither assembled genome nor available cytogenetic map.

Materials and methods

Plant material and chromosome preparation

The root tips of germinated seeds of V. unguiculata ‘BR14 Mulato’ and P. vulgaris ‘BAT93’ from Embrapa Meio-Norte (Empresa Brasileira de Pesquisa Agropecuária, Teresina, Piauí, Brazil) and Embrapa Cenargen (Embrapa Recursos Genéticos e Biotecnologia, Brasília, Distrito Federal, Brazil), respectively, were collected and pre-treated with 2 mM 8-hydroxyquinoline for 5 h at 18 °C, fixed in methanol: acetic acid (3:1 v/v) for 2–24 h at room temperature and stored at − 20 °C until use. For chromosome preparation, the tips were washed twice in distilled water, then submitted to enzymatic treatment of 2% pectolyase (Sigma-Aldrich), 4% cellulase (Onozuka or Sigma-Aldrich), and 20% pectinase (Sigma-Aldrich) at 37 °C for 2 h in a humid chamber. The slides were prepared using the air dry technique (De Carvalho and Saraiva 1993) with minor modifications. Meristems were washed in distilled water, placed individually in an approximately 30◦ vertically inclined slide. Using a Pasteur pipette, cold fixative solution (methanol:acetic acid, 3:1, v:v) were dropped over the meristem, which were macerated until its total fragmentation. Afterward, the slides were air dried by fanning, immersed in 45% acetic acid for 5 min and placed on a hot plate surface (at 37 °C) for 5–10 min.

Selection and synthesis of Oligo-FISH probes

The Oligo-FISH barcode probes were designed using the reference genome of V. unguiculata ‘IT97K-499-35’ v. 1.1 available at Phytozome (phytozome.jgi.doe.gov) and NCBI SRA BioSample accession SAMN06674009 (also ASM 411807v1) (Lonardi et al. 2019). Unique sequences of each pseudochromosome of V. unguiculata were selected after filtering for excluding repetitive sequences by Arbor Biosciences (Ann Arbor, MI, USA). Probes able to generate a signal on both V. unguiculata and P. vulgaris were selected by mapping the oligo probe derived from V. unguiculata to the P. vulgaris genome (v. 2.1, http://www.phytozome.net/commonbean.php; GenBank accession ANNZ01000000; Schmutz et al. 2014) using Blastn. Only probes having a single hit in the entire P. vulgaris genome (Blast e-value lower than 10E-05) with this hit located on the syntenic P. vulgaris chromosome were retained. Two libraries were generated, which were composed of ~ 27.000 oligos each of about 45 nucleotides long. Library 1 contained the oligo pool that generated the 16 signals detected in red after FISH, while library 2 contained the second oligo pool corresponding to the 14 green signals. These libraries covered together the thirty genomic regions selected for barcoding, which corresponded to ~ 41 megabases (Mb) of DNA sequences of V. unguiculata genome (Data S1 and S2) and were designed and synthesized by Arbor Biosciences (Ann Arbor, MI, USA). Each library was indirectly labeled with digoxigenin (library 1) or biotin (library 2) (Han et al. 2015). The complete sequence sets of both libraries are available at Supplementary data (Data S1 and S2).

Oligo-FISH and image processing

The oligo-FISH was conducted according to the protocol proposed by Braz et al. (2020b) with some modifications. The hybridization mix consisted of 50% formamide, 2 × SSC (Saline Sodium Citrate) solution (pH 7.0), 10% dextran sulfate, 350 ng of the probe labeled with biotin-green, and 200 ng of the probe labeled in digoxigenin-red, in a total volume of 10 µL per slide. The chromosomes were denatured for 5–7 min at 75 °C and incubated for 18–72 h at 37 °C in a humid chamber. After that, the coverslips were gently removed, stringency washes in 2 × and 0.1 × SSC at 42 °C (~ 76% final stringency) were conducted, followed by a wash in 1 × TNB (Tris-NaCl-Blocking) buffer. A total of 20 µL solution comprising 0.2 µL of rhodamine sheep anti-DIG (Roche) and 0.2 µL of Alexa Fluor 488 Streptavidin (Invitrogen) was applied with posterior incubation at 37 °C for 1 h. Chromosomes were counterstained with 2 µg/mL DAPI in Vectashield antifade solution (Vector Laboratories). Chromosome images were captured with Leica DM5500B fluorescence microscope, and the adjustments for brightness and contrast of images were processed using Adobe Photoshop CC (2019). The positions of each barcode marker were measured in 20 chromatids from five metaphases per species, following Fonsêca et al. (2010), except for using the DRAWID 0.26 software (Kirov et al. 2017). The new markers were integrated into the already established maps of V. unguiculata (Vasconcelos et al. 2015; Oliveira et al. 2020) and P. vulgaris (Fonsêca et al. 2010).

In silico and integrative map analysis

The sequence set for all oligos of each barcode signal was contrasted against V. unguiculata ‘IT97K-499-35’ (ASM411807v1, GenBank ID: 8,372,728) and P. vulgaris ‘G19833’ (SAMN02981484, GenBank ID: 864,298) genomes by BLASTn tool (NCBI platform), optimized for highly similar sequences (megaBLAST), for localizing them at the pseudomolecules considering e-value, score, and identity percentage variables (Table S1). Additionally, the sequences available for BAC markers used in previous studies for both species were included. BACs from V. unguiculata were previously used by Oliveira et al. (2020) and do Vale Martins et al. (2021) and are available at HarvEST:cowpea (harvest.ucr.edu). The P. vulgaris BAC sequence accessions used for the integrated map are provided in Table S2 and were obtained by BLAST using the sequences of corresponding genetic markers used to select each BAC and provided by the studies of Vallejos et al. (1992), Murray et al. (2002), Hougaard et al. (2008), and Geffroy et al. (2009). The integrated approach was represented by circular idiograms, created with Circos software v. 0.69–9 using default parameters (Krzywinski et al. 2009), enabling an in-depth comparative analysis of the cytogenetic (outer circle) and genome (inner circle) data between V. unguiculata (left side) and P. vulgaris (right side). In order to optimize collinearity, pseudomolecules, or chromosomes were eventually inverted in orientation, being the 0 Mb their initial position, as indicated in Fig. 2. The barcode signals are represented by continuous lines, and BAC signals are represented by dashed lines, each with 1 Mb of length. For both chromosomes and pseudomolecules, we used the species name abbreviations (Vu and Pv) followed by their chromosome numbering. Each V. unguiculata chromosome is represented by a different color (see Fig. 2 legend), evidencing P. vulgaris orthologous segments with the same colors. Output images (SVG format) were optimized using the CorelDraw X7 software.

Results

Oligo-FISH barcode development for V. unguiculata and P. vulgaris chromosome identification

FISH using two oligo probe sets, each containing 27,000 oligos, generated 16 red and 14 green signals on V. unguiculata metaphase chromosomes (Fig. 1a, c). Each chromosome was numbered according to the corresponding V. unguiculata pseudomolecules (Fig. 2). Thus, the barcode-FISH based on the cytogenetic map was fully integrated with the current genome sequence map. In P. vulgaris chromosomes, the two oligo-FISH probes generated 16 red signals and 13 green signals (Fig. 1b, d). The missing green signal on P. vulgaris is explained by the colocalization of the barcode signals 10A and 10B on Pv10 (Fig. 1d, f), which are separated by 3.32 Mb. Vigna unguiculata and P. vulgaris chromosomes contained one to five or four signals, respectively, which were separated by at least ~ 6.5 Mb (6.59 Mb between 4B and 4D for Pv4 and 7.65 Mb between 11B and 11C for Vu11). Except for Vu7, chromosomal arms without barcode signals had the presence of rDNA loci, as observed for Vu6, Vu10, Pv6, Pv9, and Pv10 (Fig. 2). Barcode signals covered a region from 0.19 to 1.41 Mb in length in V. unguiculata and from 0.03 to 4.07 Mb in P. vulgaris (Tables S1 and S2). Thus, it was possible to distinguish the 11 individual chromosome pairs of the two species, with each chromosome of V. unguiculata and P. vulgaris presenting a unique signal barcode pattern, which was different between species (Fig. 1).

Fig. 1
figure 1

Chromosome identification of Vigna unguiculata (2n = 22) and Phaseolus vulgaris (2n = 22) based on V. unguiculata oligo-FISH barcode. Two oligo-FISH probe sets (red and green) hybridized on mitotic metaphase chromosomes of V. unguiculata (a) and P. vulgaris (b). Homologous chromosomes in a and b were paired in karyograms to identify the 11 chromosome pairs of V. unguiculata (c) and P. vulgaris (d). Each chromosome shows a unique pattern of oligo-FISH red and/or green signals. Chromosomes were counterstained in DAPI (pseudocolored in gray). Idiograms of the barcode for both species using the V. unguiculata reference genome and considering the conserved sequences of P. vulgaris genome (e, V. unguiculata and f, P. vulgaris). Chromosomes were named according to the species name abbreviations (Vu and Pv), followed by their chromosome numbering. Each region was selected and named according to its position in V. unguiculata pseudomolecules starting at 0 Mb in alphabetical order. Bars in a–b = 5 µm and in c–d = 10 µm

Full size image
Fig. 2
figure 2

Circular representation of pseudomolecules (internal) and chromosomes (external) of Vigna unguiculata (left side) and Phaseolus vulgaris (right side) showing the position of oligo barcode, BAC, and rDNA markers. Each chromosome and pseudomolecule start at 0 Mb, indicating its orientation in the present representation. Barcode markers are represented by continuous lines, while the BAC markers by dashed lines. The color of each chromosome marker was defined in accordance with V. unguiculata chromosomes/pseudomolecules: Vu1 (dark blue), Vu2 (light green), Vu3 (red), Vu4 (gold), Vu5 (brown), Vu6 (purple), Vu7 (dark green), Vu8 (light pink), Vu9 (yellow), Vu10 (light blue), Vu11 (margeta). DNAr 5S and 35S were represented by red and green, respectively. Labels of markers that colocalize are separated by “/” and of markers that are adjacent, by “-”. The first label represents the marker that is closer to 0 Mb (from right to left in Vu and from left to right in Pv). The orientation of some pseudomolecules (Pv2, Pv3, Pv4, Pv7, Pv9, Pv10, and Pv11) and chromosomes (Pv2, Pv3, Pv4, Pv7, Pv9, Pv11, and Vu5) was inverted for a better visualization of synteny and collinearity

Full size image

Cytogenetic and in silico analysis: an integrated approach

In order to verify the differences in the barcode pattern between orthologous chromosomes and perform a more detailed comparison between species, the sequence set of each oligo barcode was mapped in its respective pseudomolecule. We integrated in silico and cytogenetic analyses, including the oligo-FISH barcode markers generated in this work and a set of previously developed V. unguiculata and P. vulgaris BAC-FISH markers, producing a total of 95 markers (Tables S3 and S4). Additionally, 5S and 35S rDNA sites were located in the cytogenetic maps.

We observed the same order for most markers, comparing both maps of the same species. However, we detected several discrepancies between cytogenetic and sequencing maps, including differences in chromosome sizes and morphologies, as well as distances between some markers. These discrepancies may be attributed to chromosome condensation differences between eu- and heterochromatic regions, low resolution of FISH in metaphase chromosomes, and to sequence gaps. Distortions in chromosome size were mainly observed for terminal 35S rDNA carrier chromosomes (Vu1, 2, 6, and 10 and Pv6, 9, and 10) because rDNA sites were not present at sequencing maps (Fig. 2). The lack of rDNA sites in pseudomolecules also altered the position of markers in short versus long chromosome arms in Vu6 and 10. To facilitate the integration of FISH and sequencing maps, chromosomes Vu4, 5, 6, and 10, as well as chromosome Pv10, were drawn in opposite orientations in relation to their respective pseudomolecules (Fig. 2).

Barcode: improvement of V. unguiculata and P. vulgaris synteny and collinearity comparisons

Based on the position of the oligo-FISH barcode markers along the chromosomes, none of the orthologous showed a conservation of the barcode pattern across the analyzed genomes (Figs. 1, 2). Chromosomes 6, 7, 9, and 11 are largely collinear, but differences in rDNA distribution between species were observed for chromosomes 6, 9, and 11.

Translocation complexes

We confirmed two translocation complexes between V. unguiculata and P. vulgaris genomes (Figs. 1, 2). Chromosomes 1, 5, and 8 were involved in a translocation complex differentiating Vigna and Phaseolus chromosomes, as previously identified by BAC-FISH, except for the participation of Pv5 (Vasconcelos et al. 2015; Oliveira et al. 2020; do Vale Martins et al. 2021), which was identified for the first time here by the presence of 1A in Pv5 and the 5A in Pv8. In silico analysis confirmed the rearrangements by the following oligo barcode markers: Vu1 short arm (Vu1S) (1A), Vu1 long arm (Vu1L) (1B, 1C); Vu5S (5B, 5C), Vu5L (5A); Vu8S (8A) and Vu8L (8B), which were located at Pv1S (8A)—Pv1L (1B, 1C); Pv5S (1A, 5B)—Pv5L (5C); Pv8S (5A)—Pv8L (8B) (Figs. 1, 2).

A reciprocal translocation involving chromosomes 2 and 3, previously identified by BAC-FISH analysis (Vasconcelos et al. 2015; Oliveira et al. 2020; do Vale Martins et al. 2021) and chromosome painting (do Vale Martins et al. 2021), was also detected using our barcode system. In the present analysis, Vu2S (2A), Vu2L (2B, 2C); Vu3S (3A, 3B), and Vu3L (3C, 3D, 3E) markers were located at Pv2S (2C)—Pv2L (2B, 3B, 3A); Pv3S (2A)—Pv3L (3E, 3D, 3C) (Fig. 1). Additional small rearrangements, mainly involving pericentromeric regions of these two chromosomes, could be evidenced in the integrated analysis, reinforcing the centromere repositioning for chromosomes 2 and 3 described by do Vale Martins et al. (2021) (Fig. 2).

Centromere repositioning of chromosomes 5, 7 and 9

In addition to the centromere repositioning for chromosomes 2 and 3 previously described (do Vale Martins et al. 2021), changes in centromere position were also detected by oligo-FISH barcode for chromosomes 5, 7, and 9 (Figs. 1, 2, 3, 4). While in Vu5 the centromere was positioned between 5A and 5B, in Pv5 it was located between 5B and 5C (Fig. 3a). For chromosome 7, the centromere was above 7A in V. unguiculata, while in P. vulgaris it was between 7A and 7B (Fig. 4a). For chromosome 9, the centromere was between 9A and 9B in V. unguiculata and between 9B and the distal BAC H10M18 in P. vulgaris (Fig. 4b).

Fig. 3
figure 3

Circular representation of the chromosomes 1, 5, and 8 of Vigna unguiculata (Vu) and Phaseolus vulgaris (Pv), showing the translocation complex among them (a), and the reciprocal translocation between chromosomes 2 and 3 (b), both identified using the oligo barcode and BAC probes. The measurement scale is presented in Mb. Each chromosome starts at 0 Mb, indicating its orientation in the present representation. Barcode markers are represented by continuous lines, while the BAC markers by dashed lines

Full size image
Fig. 4
figure 4

Circular representation of chromosomes 7 (a) and 9 (b) of Vigna unguiculata (Vu) and Phaseolus vulgaris (Pv), both suggesting a centromere repositioning. The measurement scale is presented in Mb. Each chromosome starts at 0 Mb, indicating its orientation in the present representation. Barcode markers are represented by continuous lines, while the BAC markers by dashed lines

Full size image

Inversions for chromosomes 4 and 10

Apart from changes in centromere position in chromosomes 2, 3, 5, 7, and 9, the barcode markers confirmed a pericentric inversion in chromosome 4 identified by Vasconcelos et al. (2015), resulting in a barcode pattern of green–red-green–red in Pv4 instead of green-green–red-red in Vu4, and detailed the regions involved in the breakpoints (Fig. 1e–f, Fig. 5a). This was confirmed by the sequence analysis, which revealed breakpoints between 4A and BAC 190C15 in Pv4S, and between 4B and 4D in Pv4L, or the corresponding regions in Vu4, inverting, thus, most of the chromosome (Figs. 2, 5a). We also observed a new paracentric inversion in chromosome 10, involving the segment between H025N06 and 10B in Pv10L (Figs. 2, 5b). This inversion resulted in the proximity of markers 10A and 10B in P. vulgaris, visible as a single signal in metaphase chromosomes (Figs. 1d, f, 5b).

Fig. 5
figure 5

Circular representation of chromosomes 4 (a) and 10 (b) of Vigna unguiculata (Vu) and Phaseolus vulgaris (Pv), evidencing a pericentric (a), and a paracentric (b) inversion, respectively. The measurement scale is presented in Mb. Each chromosome starts at 0 Mb, indicating its orientation in the present representation. Barcode markers are represented by continuous lines, while the BAC markers by dashed lines

Full size image

Discussion

A new chromosome identification system for both V. unguiculata and P. vulgaris was established in the present work. For the first time, the oligo-FISH barcode technique was used to identify legume chromosomes, by comparing sequence similarity of assembled genomes available for two species of related genera. This allowed us to design informative signal patterns for species belonging to different genera that diverged 9.7–15 Mya (Li et al. 2013; Garcia et al. 2020). We expanded the time of divergence achieved for sorghum and sugarcane (Meng et al. 2020), which diverged from a common ancestor for 8–9 Mya (Wang et al. 2010), and similar to the time of divergence of the six Solanaceae species evaluated by Braz et al. (2018), that diverged about 15 Mya (Wu and Tanksley 2010). Our system for beans revealed new translocation complexes, inversions, and changes in centromere position involving both crop species. Oligo-FISH barcode was also used to study chromosome evolution in crops such as potato (Braz et al. 2018), rice (Liu et al. 2020), maize (Braz et al. 2020a), and sugarcane (Meng et al. 2020), identifying rearrangements as inversions, duplications and translocations.

The bean oligo-FISH barcode allowed us to identify and compare chromosomes that could not be previously analyzed cytologically, such as Pv5, which was identified for the first time in a translocation complex that involves chromosomes 1, 5, and 8. Furthermore, a new paracentric inversion was observed for chromosome 10, in addition to the in-depth analysis of a pericentric inversion in chromosome 4, detected cytogenetically by Vasconcelos et al. (2015). Although comparative BAC-FISH studies revealed numerous rearrangements between V. unguiculata and P. vulgaris (Vasconcelos et al. 2015; Oliveira et al. 2020), there was a lack of comparative probes for chromosome 5 and some other chromosome segments, especially in the pericentromeric region (Iwata-Otsubo et al. 2016), as observed for instance for chromosomes 4 and 10. With the establishment of the oligo-based probes, it was possible to design signals for those chromosomes and segments, including signals in more proximal positions than BACs. BAC clones may contain repetitive sequences that hinder their mapping in pericentromeric, repeat-rich chromosome regions, besides being a laborious methodology. Thus, with more genome assemblies available in the last years, it is now possible to develop more efficient and versatile chromosome identification systems, expanding and accelerating cytogenetic analyses (Jiang 2019; Braz et al. 2020b). Vasconcelos et al. (2015) have described, for instance, a pericentric inversion for chromosome 4, using subterminal BACs from Pv4, while we confirmed this pericentric inversion using interstitial signals for a better characterization of the pericentromeric region of chromosome 4 of both species, narrowing down the putative breakpoint regions. Do Vale Martins et al. (2021) also reported a pericentric inversion for chromosome 4 comparing V. unguiculata and V. angularis, which belong to the subgenera Vigna and Ceratotropis, respectively, indicating that this inversion have occurred after subgenera separation (~ 3.6 Mya).

Although whole-genome comparison might be considered as the most complete way to shed light on chromosome evolution, most groups of plants still have only one reference genome, hampering comparative genomic analyses among closely related species (Varshney et al. 2012; Pecrix et al. 2018; Qin et al. 2019; Hasing et al. 2020). Additionally, rDNA loci and centromeres may be lacking in genome assemblies (Qin et al. 2019; Hasing et al. 2020). Our integrated cytogenetic and genomic approach highlighted the high level of macrosynteny and collinearity between both bean species (Lonardi et al. 2019). They revealed that some of the divergences in the barcode pattern between species were also related to changes in centromere position for chromosomes 2, 3, 5, 7, and 9. Centromere repositioning has been well documented in plants, such as cucumber and melon (Yang et al. 2014), maize (Schneider et al. 2016), and species of the tribe Arabideae (Brassicaceae) (Mandáková et al. 2020). Repositioning events can occur by successive peri- and paracentric inversion; intrachromosomal translocation; and/or acquisition of a new centromere (Schubert and Lysak 2011; Schubert 2018). For bean chromosomes 2 and 3, a complex rearrangement was confirmed. Besides a major translocation event, minor rearrangements were observed, such as inversions, intrachromosomal translocations, and centromere repositioning by new centromere, since collinearity was not altered between Pv2 and Pv3 pericentromeric region, as discussed by do Vale Martins et al. (2021).

For chromosomes 5, 7, and 9, centromere repositioning can be due to inversions, intrachromosomal translocation, and/or new centromere, although an association of the translocation complex involving chromosome 5 cannot be excluded. More markers are necessary to understand the involved mechanism for each of these centromeric changes. A recent study comparing P. lunatus (lima bean) and P. vulgaris genomes identified a complex intrachromosomal translocation within chromosome 9 (Garcia et al. 2020), which may be associated with the centromere repositioning observed between both species for this chromosome (Bonifácio et al. 2012). Further genome comparisons are necessary to confirm if this event also explains the difference observed between P. vulgaris and V. unguiculata chromosome 9, what, in this case, would suggest that V. unguiculata and P. lunatus chromosomes might represent the ancestral state. New centromere formation is common in domesticated plants, such as maize and potato (Talbert and Henikoff 2020), and is closely related to its domestication time (Schneider et al. 2016). This phenomenon may be associated with the selection of genes linked to the centromere. Thus, Vigna and Phaseolus beans, with multiple domestication events each, might be great targets to understand if domestication favors centromere repositioning.

Thus, the oligo-FISH barcode presented in this study provided the first legume chromosome identification enabling to distinguish individual chromosome pairs, besides identifying chromosome rearrangements and centromere repositioning not described previously. This technique can be applied to other crops of both genera that lack genomic information, such as moth bean (V. aconitifolia), hairy cowpea [Vigna luteola (Jacq.) Benth.], wild cowpea (Vigna vexillata L.) yearlong (Phaseolus dumosus Macfad.), scarlet runner (Phaseolus coccineus L.), and tepary bean (Phaseolusacutifolius A. Gray). It will probably also be useful for infrageneric comparisons of related genera, thus helping to understand the chromosome evolution of this important group of legumes.