Meiosis and male fertility in F1 interspecific hybrids (Passiflora vitifolia vs. Passiflora hatschbachii)

Interspecific hybrids can be studied using methodologies in which the male gamete with high reproduction potential, viability, and fertility is prioritized. Passiflora species, with lush, showy, and exotic colors, have great potential for the ornamental plant market. In addition, artificial Passiflora hybrids were developed without many difficulties because of weak reproductive barriers. Thus, meiotic and post-meiotic behaviors were analyzed with 2% acetic carmine staining. Confirmation of interspecific hybridization was performed using SSR markers and GISH technique was used to detect genomic differentiation. The pollen viability of the parental and hybrids genotypes was tested using Alexander`s solution, fluorescein diacetate and in vitro germination tests were performed using culture medium. The meiotic behavior was regular and displayed haploid number n = 9 with nine bivalent (II) chromosomal, and pairing in 90% of the cells in diakinesis. There was a significant difference (p < 0.05) in terminal and interstitial chiasma frequencies. Meiotic irregularities observed were as follows: early and/or delayed chromosomes, disorientation of spindle fibers, transverse spindles, tripolar spindles, and asynchrony; and consequently irregular post-meiotic products were observed: monads, dyads, triads, and polyads. GISH was used in the interspecific hybrids and pairing between homeologous chromosomes, and bivalent and tetravalent formation were observed. From this study, we could conclude that hybrid genotypes are fertile and pollen grains are viable and can be used in breeding programs.


Introduction
The Passifloraceae family (A. L. de Jussieu ex. Kunth) belongs to the order Malpighiales, encompassing approximately 17 genera and more than 700 species (Bernacci et al. 2013). The genus Passiflora L. is the largest in terms of number of species, that is, more than 525 species (Cervi and Imig 2013), distributed in the tropical and subtropical regions (Ulmer and MacDougal 2004). Brazil is regarded as the center of origin and geographical distribution of a large number of species of this genus (Meletti et al. 2000;Souza et al. 2003). These species present high Abstract Interspecific hybrids can be studied using methodologies in which the male gamete with high reproduction potential, viability, and fertility is prioritized. Passiflora species, with lush, showy, and exotic colors, have great potential for the ornamental plant market. In addition, artificial Passiflora hybrids were developed without many difficulties because of weak reproductive barriers. Thus, meiotic and postmeiotic behaviors were analyzed with 2% acetic carmine staining. Confirmation of interspecific hybridization was performed using SSR markers and GISH technique was used to detect genomic differentiation. The pollen viability of the parental and hybrids genotypes was tested using Alexander`s solution, fluorescein diacetate and in vitro germination tests were performed using culture medium. The meiotic behavior was regular and displayed haploid number n = 9 with nine bivalent (II) chromosomal, and pairing in 90% of the cells in diakinesis. There was a significant 1 3 2 Page 2 of 19 Vol:. (1234567890) levels of genetic diversity (Meletti et al. 2005;Silva et al. 2014) and are of economic interest; they can be used for pharmacological, nutritional and cosmetic purposes, and are of the ecological and ornamental interest (Abreu et al. 2009;Ocampo et al. 2016).
Passiflora species have great potential for the ornamental plant market because the flowers exhibit exuberant, showy, and exotic colors (Abreu et al. 2009). The morphology of the flowers, especially the androgynophore, and the complex corona, consisting of filaments arranged in one or more concentric rows, are beautiful. Also there is interspecific variation in size, shape, and color, which hare striking features for the genus (Vanderplank 2002). However, in Brazil, the use of these species for ornamental purpose is minimal, while in Europe these species are used for the decoration of pergolas, walls, and gardens (Junqueira et al. 2008;Abreu et al. 2009;Conceição et al. 2011). In addition, artificial Passiflora hybrids can be developed without many difficulties (Vanderplank 2000) because of weak reproductive barriers (Belo et al. 2018;Souza et al. 2020). Several crosses involving Passiflora species have shown that the success in obtaining hybrids is directly related to the chromosome number (Conceição et al. 2011;Santos et al. 2012;Belo et al. 2018;Souza et al. 2020). When crosses are carried out between species with different chromosome numbers, there is no success (Conceição et al. 2011). This is probably associated with the fact that species with the same chromosome number share many repetitive DNA sequences .
Interspecific hybrids of Passiflora with interesting features can be developed as a strategy to create varieties for the ornamental plant market. Interspecific hybrids are sources of variability, which may allow the selection of promising genotypes for plant breeding. Hybrid genotypes usually give a higher flower that yield and provide a wider variety of shapes and colors (Conceição et al. 2011). Ensuring the success of interspecific hybridization through regularities of meiotic events such as pairing, even partially, between the homeologous chromosomes, chromosomal recombination, regular meiosis, and viable pollen grains (PG) are important (Souza et al. 2003). Meiosis depends on the regularity of events such as chromosomal pairing of bivalents, formation of the synaptonemic complex, occurrence of crossing over and chiasmas, recombination, and formation of regular post-meiotic products. Regular meiotic behavior enables the formation of viable gametes (Belo et al. 2018), and thus, the formation of fertile hybrids (Lavinscky et al. 2017(Lavinscky et al. , 2021. Time and cost reduction are interesting factors for improvement programs. The costs of plant maintenance can be reduced by confirming the hybridization still in seedling phase, and molecular markers have been widely used for this purpose (Bellon et al. 2007;Junqueira et al. 2008;Conceição et al. 2011), in addition to genomic in situ hybridization (GISH) techniques Souza et al. 2020). Molecular markers are also used to verify the occurrence of DNA polymorphism (Fajardo et al. 1998) once the markers generate information on allelic diversity of the analyzed groups and phylogenetic relationships . The simple sequence repeats (SSR) are short nucleotide sequences (1-10 bp), repeated in tandem, occurring in prokaryotes and eukaryotes Vieira et al. 2016). Sequences that flank the microsatellites are conserved even among species, which are used for drawing primers and amplification of SSR (Goetze et al. 2013). Studies have been already performed in which SSR were used to confirm interspecific hybridization between species of the genus Passiflora (Melo et al. 2016). In addition, successful assessment of the transferability rates of SSR primers in wild species of Passiflora has been made (Silva et al. 2014).
GISH can be used to confirm interspecific crossings ; this technique can be explored by distinguishing the parental genomes in hybrid genotypes, differentiating the chromosomal lots from the supposed parents Younis et al. 2015;Ramzan et al. 2017). The use of GISH can be broadened, that is, the technique can be used to assess the phylogenetic inferences between species with genomic homology Silva and Souza 2020). GISH technique also can be used to observe the occurrence of introgression of genes of interest in the genetic improvement of plants (Ramzan et al. 2017), beyond confirming hybridization produced by apomythic species and allopolyploid species (Younis et al. 2015). In reports based on the study of meiotic behavior, it has been revealed that the GISH technique assists in the analysis of chromosomal matching between homologous and homeologous chromosomes (Xie et al. 2010(Xie et al. , 2014. It is possible, using GISH, to observe the formation of Page 3 of 19 2 Vol.: (0123456789) multivalent, bivalent, univalent chromosomes, and chiasmas frequency among the homeologous chromosomes (Younis et al. 2015). GISH can also be used to obtain important information about factors that cause meiotic irregularities and how these factors affect fertility in plants . Thus, in this study, we aimed to (i) produce interspecific hybrids of Passiflora for ornamental use; (ii) confirm hybridization with the use of molecular markers and GISH; (iii) analyze meiotic behavior, pollen viability of the parental species, and hybrid genotypes.

Plant material and Interspecific hybridizations
Passiflora species used for artificial interspecific crossing in the present study were maintained at the Passiflora work collection, located at the Universidade Estadual de Santa Cruz (UESC), in Ilhéus, Bahia, Brazil (39° 10″ W, 14° 39″ S, 78 m a.s.l.). Artificial crossings were performed between species P. vitifolia Kunth (accession 481) vs. P. hatschbachii Cervi (accession 486) from October 2015 to January 2016, during the early hours of the day, that is, between 7:00 and 8:00 a.m. in the anthesis. Flowers of P. vitifolia and P. hatschbachii were used as pollen recipients and donors, respectively. On the previous day, pre-anthesis flower buds of the maternal parent P. vitifolia were emasculated and protected using paper bags. PG from the paternal parent (P. hatschbachii) were collected in bulk and the flowers with receptive stigma from the maternal parent, already curved, were pollinated manually, and protected again using paper bags. The fruits were identified and protected with nylon nets until complete maturation.

Cultivation conditions
The collected seeds were placed in Promalin® for 72 h to break dormancy. In all, 202 seeds were placed in plastic trays containing a commercial substrate. Sixty-seven hybrid seedlings were transplanted and kept in 1.5 L polyethylene bags, containing substrate, and in 50% shade in a green house. After 90 days, hybrid plants were brought to the field in and kept in sunlight. All plants obtained from the interspecific crosses were named as progeny HD26. The analyzes were carried out in the parents and eight hybrid genotypes that were randomly chosen: HD26-104, .

Extraction of genomic DNA and SSR amplification
The total genomic DNA was extracted from fresh and young leaves from parents and putative hybrids using a CTAB-based protocol (Doyle and Doyle 1990) with some modifications (Viana et al. 2006). DNA quantification was done, and quality was assessed using a Nanodrop DN-1000 spectrophotometer. We performed marker SSR cross-amplification tests of primers originally developed for P. alata  and P. edulis (Oliveira 2008) (Table 1) in the parents and putative hybrids. Amplification reactions were performed in a total volume of 20 µL containing 20 mM Tris-HCl, 50 mM KCl, 2.5 mM MgCl 2 , 2.5 mM of each desoxinucleotides (dATP, dTTP, dGTP, and dCTP), 10 pmol of each primer, one unit of Taq polymerase enzyme, and 20 ng/μL of DNA. The amplification conditions were as follows: initial denaturation at 95 °C for 5 min, followed by 38 denaturation cycles at 94 °C for 1 min, annealing temperature of 56 °C for 40 s, extension at 72 °C for 50 s, and a final extension at 72 °C for 7 min. Agarose gel (1% stained SYBR®) was used to check amplification profiles, using TBE 1 × and the molecular marker Ladder 100 bp to observe the size of the amplified DNA fragments, allowing confirmation of hybridization.

Meiosis and post-meiosis
Flower buds at different stages of development were collected in the morning. The buttons were placed in Carnoy solution [ethanol:glacial acetic acid (3:1, v/v)] at room temperature (RT) (Johansen 1940) with exchange of Carnoy being carried out at 30 and 180 min and stored at − 20 °C for 24 h. The buds were transferred to 70% ethanol and kept at ± 10 °C until analysis. For the observation and counting of meiotic phases and post-meiotic products, one anther from each button was analyzed, and temporary slides were prepared using the squash technique and stained with 1% acetic carmine (Souza and Pereira 2011). Fifty cells of each genotype were observed for each meiosis phase. The frequencies of terminal, interstitial, and total chiasmas were verified by calculating the recombination index (RI = [Σn° total number of chiasmas ÷ n° of cells analyzed] + n value) (Darlington 1958). We considered the occurrence of one chiasma for the rod-paired bivalents (rod) and two chiasmas for the ring-paired bivalents (ring) (Senda et al. 2005). The meiotic index (MI (%) = [n° of normal tetrads × 100] ÷ n° of post-meiotic products counted) (Love 1951) was calculated from the number of postmeiotic products (monads, dyads, triads, tetrads, and polyads) and for analysis of variance (ANOVA), considering a completely randomized design, with replicates of four slides per genotype, the SISVAR opensource computer program was used.
In order to register meiotic irregularities, slides of meiotic phases were prepared in accordance with the report by Pierre et al. (2011), for the removal of the callose and staining with DAPI. Adaptations were made using this methodology: the anthers were washed twice using distilled water, for 5 min each time, by agitating, and macerated in the enzyme pectinase (Sigma); and the Eppendorf tubes with the macerated anthers were incubated at 37 °C for 20 min and then centrifuged for 10 min-2000 rpm to obtain suspension cells. The supernatant was discarded and centrifuged for 10 min at 2000 rpm with distilled water. Again, the supernatant was discarded and 20 µl of 45% acetic acid was added to obtain the cell suspension, which was dripped on the slides, covered with 20 × 20 mm coverslips, and then placed in liquid nitrogen for 15 min for the fixation of the material on the slides. Furthermore, the slides were dried at RT for another 15 min and stained with 2 μL DAPI (2 μg/mL; Sigma) and detected with a U-MWU filter (330-385 nm excitation/emission > 420 nm) and photographed under an Olympus BX41 epifluorescence microscope equipped with a 5 M Olympus DP25 digital camera and DP2-BSW software.

Preparation of probes for genomic in situ hybridization (GISH)
For genomic differentiation using the GISH technique, slides were made with the use of anthers in the diakinesis subphase (prophase I) and metaphase I phase, and the procedure for the preparation of slides was the same as that used for meiosis with irregularities, as described in the previous section. To make the probe, the genomic DNA of the parental species was fragmented using a sonicator (QSonica), through programming: 40% amplitude, 2 s on and 2 s off, in a total period of 5 min each (Jauhar and Peterson 2006), and the fragment size was verified by electrophoresis in 2% agarose gel (Pronadisa) using a 100 bp ladder as the marker (New England Biolabs) and stained with SYBR safe (Invitrogen), with the preference of using cleaved fragments between 200 and 1000 bp. The probes were marked by Nick Translation following the protocol proposed by the manufacturer, labeling the paternal parent probe (P. hatschbachii) with Biotin-16-dUTP (Roche Diagnostics) and maternal parent probe (P. vitifolia) with digoxigenin-11-dUTP (Roche Diagnostics), with a final concentration of 1 μg of cleaved DNA from each parent.

GISH in meiosis
The GISH technique followed the protocol proposed by Schwarzacher and Haslop-Harrison (2000) with modifications by Melo et al. (2015). Four slides with good cytological preparations for each genotype were dried at 37 °C for at least 1 h. Following treatment with 50 μL of a solution containing 100 μg/mL RNase (Sigma-Aldrich) in 2 × SSC (saline-sodium citrate buffer; 0.3 M sodium chloride [Sigma] and 0.03 M sodium citrate [Sigma]), the slides were incubated in a humid chamber at 37 °C for 1 h. The slides Page 5 of 19 2 Vol.: (0123456789) were washed twice in 2 × SSC for 5 min each. 50 μl 10 mM hydrochloric acid (HCl; Vetec) were added over the slides for 5 min. Following this, HCl was removed and replaced with 50 μl of pepsin (Sigma) [10 mg pepsin/ml, 10 mM HCl (1:100 v/v)] and the slides were incubated in a humidified chamber for 20 min at 37 °C. Then, the slides were washed twice in 2 × SSC for 5 min, immersed in 4% paraformaldehyde (Sigma) at RT for 10 min, and washed twice in 2 × SSC for 5 min. The slides were dehydrated in 70% and 95% ethanol for 5 min each and dried at RT for 30 min. The hybridization mixture (50% formamide, 10% dextran sulfate, 2 × SSC, and 0.13% sodium dodecyl sulfate) was added at a final volume of 15 μL. For the hybridization mixture, 33 ng of each DNA probe parent was added, which was heated at 75 °C for 10 min in a thermocycler (Eppendorf, Mastercycler) and placed on ice for 5 min. The slides with the hybridization mixture were denatured in a thermocycler with a slide adapter (Techne, TC-412) at 75 °C for 10 min and incubated overnight at 37 °C in a humidified chamber. After hybridization, slides were immersed in 2 × SSC for 5 min at RT to facilitate coverslip removal, moved to a Dubnoff bath (Marconi, MA093/1/E) set at 42 °C, and immersed in 2 × SSC for 5 min each, twice in 0.1 × SSC for 5 min each, and twice again in 2 × SSC for 5 min each. Finally, slides were dipped in 4 × SSC containing 0.2% Tween 20 (Sigma) at RT for 5 min and then treated with 50 μl of 5% bovine serum albumin (BSA; Sigma). Biotinlabeled probe was detected using 0.7 μL of avidinfluorescein isothiocyanate (FITC; Vector), and the digoxigenin-labeled probe was detected using 0.7 μL of anti-digoxigenin-rhodamine (Roche Diagnostics) plus 18.6 μL of 5% BSA solution per slide. Slides containing the antibody for detection were incubated in a humid chamber at 37 °C for 1 h, and three baths were performed in 4 × SSC/0.2% Tween 20 in RT for 5 min each to remove excess antibodies. The slides were quickly immersed in 2 × SSC, mounted, and counterstained with DAPI/Vectashield (Vector H-1200), and stored at 8-10 °C until analysis.

Pollen viability and in vitro germination
To perform pollen viability and in vitro germination tests, flowers were collected randomly at the beginning of the anthesis in the morning. PG corability was tested with (i) fluorescein diacetate (FDA) (Sigma) (Heslop-Harrison and Heslop-Harrison 1970), fluorochrome which indicates the presence of esterase enzyme activity related to the integrity of the plasmalemma, and (ii) Alexander`s solution (Alexander 1969), which used triple staining with Orange G (intensifier), basic Fuchsin (stains the cytoplasm red; Sigma) and Malachite green (stains the PG wall green) which provided the reactivity of the wall/cytoplasm. For FDA, the slides were prepared by adding a drop of glycerol to the FDA and incubated for 10 min in a humid chamber at 37 °C. PG whose cytoplasm were stained greenish yellow fluorescent were considered viable. The PG were counted and observed for fluorescence colorability using the Olympus BX41 epifluorescence microscope with U-MWB filter (450-480 nm excitation/emission > 515 nm), equipped with an Olympus DP25 5 M digital camera and DP2-BSW software. We used four different button anthers of each genotype (four repetitions), totaling 2000 PG counted by genotype. Alexander`s solution was used to differentiate the PG into viable (with stained and intact cytoplasm) and unviable, which were classified as T1 (contracted cytoplasm); T2 (absence of cytoplasm-empty); T3 (micropollen), and T4 (giant). We used four different button anthers of each genotype (four repetitions), totaling 4000 PG counted by genotype. The results were submitted to ANOVA with the aid of the open-source computer program SISVAR.
For in vitro germination of pollen, culture medium was used by following the protocol proposed by Bruckner et al. (2000), with modifications. The PG from each anther were put in a slide with a drop of autoclaved medium (0.10 g/L of boric acid (H 3 BO 4 ; Merck), 50 g/L of sucrose (Merck), and 0.3 g/L calcium nitrate tetrahydrate (Ca(NO 3 ) 2 . 4H 2 O; Merck); 0.2 g/L magnesium sulfate heptahydrate (MgSO 4 . 7H 2 O; Merck), and 0.1 g/L of potassium nitrate (KNO 3 ; Merck)) and incubated in a humid chamber at 37 °C for 24 h. Germinated GP were counted by genotypes totaling 500 GP per slide. The results were submitted to Skott-Knott test with SIS-VAR (Ferreira 2011).

Interspecific crossing confirmation via SSR
Of the tested primers, Pa07 amplified the polymorphic locus in the parents, confirming the hybrid character of the F 1 hybrid (P. vitifolia vs. P. hatschbachii) progeny. The DNA amplification products showed an informative band of the maternal parent of approximately 280 bp, and another informative band of the paternal parent was about 257 bp. The heterozygote profile with bands inherited from the genitor species can be observed in hybrids (Fig. 1).

Meiotic behavior
The parents and hybrids progeny analyzed presented haploid number n = 9 and regular meiosis. In chromosomal pairing, nine bivalents (II) were observed in 90% of diakinesis cells. The species used in the interspecific hybridization in this study, P. vitifolia and P. hatschbachii have high cariotypical compatibility, with a chromosomal number of 2n = 18; this made it possible to observe a high frequency of bivalent chromosomes in F 1 hybrids (above 90%). It were analyzed interspecific hybrids and observed regular meiotic behavior with chromosomal alignment in the metaphase plate, anaphase with regular segregation, and formation of four haploid nuclei at the end of telophase II (Fig. 2a-h) in 90% of the cells. Regarding the chromosomal pairing configuration, the formation of bivalents in diakinesis was greater than 90%, both in parents and in hybrids, but it was observed some irregular chromosomal pairing (Table 2; Fig. 3).
Terminal and interstitial chiasmas were observed in all genotypes, except in the HD26-152 hybrid that did not present terminal chiasmas and resulted in the highest mean of interstitial chiasmas. RI ranged between 9.3 (HD26-118) and 11.1 (HD26-146) ( Table 3). The analysis of variance showed a significant difference (P < 0.05) in the frequencies of terminal and interstitial chiasmas between hybrid genotypes and parents (Table 4). In this study, there was occurrence of interstitial chiasmas in all genotypes examined. In the HD26-152 genotype, we could not visualize terminal chiasmas; however, recombination was not impaired, because it is the interstitial chiasmas that guarantees recombination between chromosomes.
Irregularities in meiotic behavior were also observed during meiosis I and II (Fig. 2i-k; Table 5), such as early and/or laggard chromosomes, spindle irregularities and asynchrony (Fig. 4). Some meiocytes presented chromosomal bridge formation during anaphase II segregation (Fig. 4f, h). Spindle irregularities were found in all genotypes in telophase II, and transverse spindles (chromosomal sets positioned on the equatorial plate in "T" shape) and tripolar spindles ("V" shape) were observed in meiocytes in metaphase and anaphase II (Fig. 4h, i). In meiosis II, asynchrony was the irregularity least found among genotypes (Table 5), with meiocytes presenting a chromosomal group in prophase II, while the other was in metaphase II, or one chromosome group in metaphase II and the other initiating anaphase II and meiocytes with one chromosomal group in anaphase II and the other beginning and/or end of telophase II.
The hybrids progeny and parental Passiflora evaluated presented a meiotic index (MI) above 90%, which can be considered to represent cytological stablity and the hybrids being potentially fertile. The analysis of post-meiotic products showed normal tetrad in more than 90% of genotypes (Table 6, Fig. 2l). However, irregular post-meiotic products such as monads, dyads, triads, and polyades were also found ( Table 6, Fig. 2m-o). Triads were the most frequently found irregularities in all genotypes (Fig. 2m), while other irregularities did not reach 0.5% of the analyzed cells. The micronuclei observed in this study probably originated as a result of the metaphases with early chromosomes and laggard chromosomes in the anaphase (I and II).

GISH
Interspecific hybridization was confirmed with the use of the GISH technique, which was applied in the diakinesis phases of prophase I and at the beginning of metaphase I. It was possible for us to distinguish the genome of the parent species in the hybrid genotypes and visualize pairing between homeologous chromosomes in the HD26 progeny (n = 9), with the formation of bivalents (Fig. 5). In HD26-143, homeologous chromosomes paired on the metaphasic plate and connected to the spindle fiber were observed, showing regular meiosis (Fig. 5a). In HD26-146, tetravalent formation was observed among homeologous chromosomes (Fig. 5b).

Pollen viability and in vitro germination
The PG reacted positively to the histochemical test with the dyes used, Alexander`s solution, and fluorescein diacetate (FDA), both in the parents and HD26 progeny hybrids (Fig. 6). Alexander test was useful in enabling the observation of viable PG with intact wall and cytoplasm, in addition to the four types of unviable PG. The percentage of viable PG was 65.5% and 96.5% in the P. vitifolia and P. hatschbachii parents, respectively. In hybrids, the highest percentage of viable PG was observed in the HD26-136 genotype   Fig. 6). The analysis of variance showed a significant difference (P < 0.05) in pollen viability with Alexander`s solution and for the different types of unviability, T1-T4 (Table 8). With regard to FDA staining, the analysis of variance showed no significant difference between analysed variables (viable and unviable pollen grains) (Table 8), however the use of this technique was efficient, revealing viable fluorescent yellow-green PG with percentage values from 34.3 (parental) to 95.6 (hybrid). For in vitro germination, the paternal parent obtained the highest percentage of germinated PG (62.3%) (Fig. 5b) and among hybrid genotypes, HD26-137 showed the very low percentage (11.2%). In general, hybrid genotypes presented a low percentage of pollen germination, and there was no germination for some hybrid genotypes analyzed (Table 9). The floral anthesis of hybrid genotypes and maternal parent (P. vitifolia) occurs around 7.30 a.m. in the morning. For the reason that collection was done in the winter, during rainy days, collection was done in the morning between 9:00 and 10:00 a.m. Hence, the collection occurred after anthesis of the genotypes, because of which unviable PG were observed in histochemical tests, and which also have affected in vitro germination. The high percentage in the in vitro germination test found for parent P. hatschbachii corroborated with the high pollen viability shown with Alexander`s solution. We observed a significant difference between genotypes, evaluated by the Skott-Knott average test, with in vitro pollen germination.

Discussion
The interspecific crossing between P. vitifolia vs. P. hatschbachii resulted in the hybrid progeny HD26 with haploid number n = 9, similar to the parents, and with regular meiotic behavior. The weak barriers among species with the same chromosome made it possible to obtain hybrids involving these two species (Conceição et al. 2011). Several interspecific hybrids of Passiflora have been successfully obtained (Conceição et al. 2011;Santos et al. 2011;Belo et al. 2018;Souza et al. 2020;Lavinscky et al. 2021).
The molecular marker SSR confirmed interspecific crossing with the presence of specific band of each parent in hybrid genotypes. As a highly polymorphic, co-dominant marker, SSR allows differentiation between homozygous and heterozygous individuals (Zanella et al. 2017). Compared to other molecular markers, SSR is easily visualized and stable (Song et al. 2005). Furthermore, phylogenetically close species share conserved DNA sequences, and by homology, the transferability of primers in SSR regions in hybrid genotypes is facilitated (Kalia et al. 2011). In this study, we tested SSR primers drawn for P. edulis f. flavicarpa Deg. and P. alata Curtis. We obtained only amplification primers for P. alata, while other studies of Passiflora with primers of P. edulis f. flavicarpa were successful, with high transferability (Belo et al. 2018). However, amplification of a few primers can confirm interspecific crossings. A specific locus of each genitor present in the hybrid progeny was sufficient to confirm hybridization (Conceição et al. 2011;Souza et al. 2012;Melo et al. 2016). HD26-146 with homeologous chromosomes paired at diakinesis and one tetravalent (arrow). Bar = 10 μm In the hybrids progeny analysed we observe chromosomal pairing and formation of bivalents. The formation of bivalents indicates structural and genomic similarity among the parent species (Abreu et al. 2009;Lavinscky et al. 2021), providing evidence of gene flow between this species. The pairing of homeologous chromosomes in hybrids increases the possibility of new allelic combinations and is directly related to the viability/fertility of hybrids (Abreu et al. 2009). In interspecific hybrids, the absence of pairing between the different parental genomes in the meiosis phases may compromise the reproductive potential of interspecific hybrids. In addition, the genetic distance between the species involved in the crossing may lead to irregularities of linkage in meiosis, as was observed in the crossbreeding hybrids of Orchdaceae Juss., Paphiopedilum delenatii Guillaumin × P. callosum (Rchb.f.) Stein and P. delenatii × P. glaucophyllum J.J.Sm. (Lee et al. 2011).
In this study, the parental species and the hybrids progeny presented more than 90% of cells in diakinesis with paired bivalent chromosomes. But we also found different chromosomal association such as univalent, trivalent, and tetravalent (evidence for structural change) which suggest similar genetic structure between the species involved in the crossing . Different chromosomal association was observed in interspecific hybrids of crossing P. gibertii vs. P. gardneri (Lavinscky et al. 2021), which led the formation of precocious and lagging chromosomes in all phases of meiosis. Univalent formation is directly related to low frequency chiasmas, which can affect pollen viability (Pagliarini 2000;Souza and Pereira 2011). However, chromosomal configurations such as univalent and multivalent are frequent in Passiflora hybrids (Soares-Scott et al. 2003;Souza and Pereira 2011;Lavinscky et al. 2021). The average number of terminal, interstitial and total chiasmas observed in cells in the diakinesis in this study was high compared to other species of Passiflora, the mean of total chiasmas per cell in the progeny hybrids of the crossing P. gibertii vs P. gardneri and parents ranged from 5.3 to 9.14 (Lavinscky et al. 2021). Terminal chiasmas ensures perfect disjunction as well as segregation of homologous chromosomes and maintain the physical structure of the bivalents during prophase I, and not interfering with recombination index . Other studies on species Passiflora and hybrids corroborate high index of interstitial chiasmas (Souza et al. 2003;Souza and Pereira 2011), as it was found here. Indeed, interstitial chiasma is an indicator of recombination (Lavinscky et al. 2017). The average frequency of chiasmas per bivalent has relevance in meiosis stability (Pagliarini 2000), and the average in our study can be considered high for Passiflora species, which is an important aspect that can prevent premature separation of bivalents, resulting in univalent chromosomes (Souza and Pereira 2011).
The meiotic irregularities found here commonly occur in Passiflora species (Souza et al. 2003Pereira et al. 2017;Lavinscky et al. 2021) such an asynchronism, chromosomal bridges and disorientation of spindle fibers that can be caused by the depolymerization of spindle fibers, leading simultaneously to two chromosomal sets being in different stages of meiosis (Souza et al. 2003). As well as asynchronism occur because these chromosomes do not reach the telophase phase for regular chromosomal segregation (Soares-Scott et al. 2003;Corrêa et al. 2005;Damasceno et al. 2010;Kiihl et al. 2011;Souza and Pereira, 2011). In addition, paracentric translocations and inversions can cause the emergence of chromosomal bridges (Levin 2002;Paiva et al. 2012). In studies based on cereals, it was reported that sticky chromosomes may have caused the origin of chromosomal bridges in the phases of anaphases (Paiva et al. 2012).
The meiotic index between 90 and 100% give the plants a stable cytology confidence (Love 1951). In our study meiotic indices of 90% represent cytological stability, indicating that cells with meiotic irregularities were less frequent than normal cells, suggesting the efficacy of checkpoints during meiosis. The main irregular post-meiotic products such as triads may be caused by the occurrence of irregularity spindle fibers (Souza and Pereira 2011) as well as errors in cytoplasmic division during meiosis I and II (Moreira et al. 2017). In addition, micronuclei can also be formed by cytomixis (transfer of material from one cell to another) (Diegues et al. 2015), but this event was not visualized in this study. Postmeiotic products with irregularities can originate from gametes with an extra chromosomal set, nondegraded gametes, or generate aneuploid gametes (Souza et al. 2003), or unbalanced PG (Moreira et al. 2017) or sterile .
The GISH technique confirmed the interspecific crossing between the parental species P. vitifolia and P. hatschbachii applied to meiocytes in the diakinesis subphase (Prophase I) and in metaphase I as well as allowing the visualization of chromosomal pairing. Although the diakinesis subphase in prophase I is the ideal stage for viewing chromosomal pairing because of the difficulties encountered in eliminating or decreasing physical barriers in meiocyte (cell wall, calosis, and cytoplasm) cells were also used in metaphase I. In studies based on banana (Musa L.), the same difficulties were observed with regard to breaking of the cell wall and the cytoplasmic density (Capdeville et al. 2009;Jeridi et al. 2011); these physical barriers hinder the probe of chromosomes (Jeridi et al. 2011), requiring an adequate number of protocols.
In this study, using GISH, it was possible for us to observe the pairing between homeologous chromosomes as well as pairings between chromosomes of the same genome. Plant genomes share many repetitive DNA regions, and in Passiflora, this type of DNA is part of the genome Silva and Souza 2020). Homeologous chromosome pairings were observed in all hybrid genotypes evaluated, confirming the existence of homology between genomes, which consequently allows recombination between chromosomes of different species. GISH on meiotic chromosomes was used to confirm polyploidy in Dahlia Cav. by interspecific crossing, which formed bivalents in cell pairing between parental genomes in the metaphase I; homeologous and homologous chromosomes were observed in the formation of tetravalents, trivalents, and univalents (Gatt et al. 1999). In this study, the formation of tetravalents and multivalents was observed. Interspecific hybrids of Musa acuminata (AA) and M. balbisiana (BB) were also found to be univalent, bivalent, trivalent, and multivalent among homeologous chromosomes (Jeridi et al. 2011). The use of two probes simultaneously for the differentiation of parental genomes in hybrid genomes was satisfactory in this work; the application of this methodology has already been reported to be successful in other studies, which, in addition to differentiating parental genomes in interspecific hybrids in Passiflora, has also allowed the observation of recombinant chromosomes in backcrossed hybrids Souza et al. 2020).
Pollen viability is directly related to regular meiosis. Plants with regular meiotic behavior present viable PG , and it is important to measure the potentiality of fecundation and fertilization of the PG (Biondo and Battistin 2001;Soares et al. 2013). A high percentage of viable PG is related to the high index of normal tetrads, a result of regular meiotic behavior . In this study, the histochemical test with Alexander solution showed some hybrid genotypes with high pollen viability percentages, with 70% approximately. For Passiflora species, percentage value above 70% is considered high (Souza et al. 2002). Pollen viability can be favored by several factors, such as temperature variation, humidity, time, and storage conditions of the PG (Franzon et al. 2005). In addition, in Passiflora, high viability is favored because the morphology of the PG presents a substance known as pollenkitt that offers a sticky consistency to the PG, serving as protection, allied to physiological factors avoiding dehydration of the PG (Souza et al. 2002). In some species, the correlation between the loss of pollen viability and water loss has been reported (Zanatto et al. 2009). Pollen viability with an average of less than 30% is considered low, while between 30 and 69% is considered moderate. In this study, the hybrid genotype with the highest percentage of pollen viability was 84% in HD26-136. In studies with interspecific pepper hybrids of the genus Capsicum L., the pollen viability of 72.5% was considered satisfactory (Moreira et al. 2017).
Over time, after flowering, pollen viability may be compromised, with the tendency to decrease. Variations throughout the day, such as temperatures, possibly cause the PG to become unviable (Belo et al. 2018). In addition to abiotic factors that may interfere (Abreu et al. 2009), pollen viability may vary among genotypes of the same species, cultivars, flowers of the same plant, anthers in the same flower, among other factors (Souza et al. 2002;Coelho et al. 2012). Studies about pollen feasibility that dynamically shows the changes that occur during flowering stages in a plant are important to assist in the characterization of reproductive barriers and in the selection of fertile PG (Deng et al. 2017). Floral development studies were performed by Soares et al. (2018), who evaluated Passiflora species, with regard to pollen viability, stigma receptivity, and in vivo and in vitro germination of PG, in the three stages of flowering: pre-anthesis, anthesis, and post-anthesis. The results show the ideal moment of PG fertilization potential and consequently the fructification is in the anthesis for the vast majority of the species analyzed. This moment of PG fertilization is fundamental to overcome pre-fertilization and post-fertilization barriers at crossings (Soares et al. 2018).
Contrary to what has been seen for the maternal parent (P. vitifolia) and some hybrid genotypes that had pollen viability and in vitro germination in high humidity, the humid environmental condition may have allowed water retention in the PG in P. hatschbachii and its viability may have been extended for a longer time after anthesis, maintaining its optimum temperature for germination (Souza et al. 2002). The anthesis of this species occurs early in the morning and the flowers remain open until dawn in optimum temperature conditions. Thus, environmental factors may influence pollen viability and consequently pollen germination, as was observed for the parent species in our study. Other tests of viability and germination of PG in Passiflora species are recommended, as there are few studies on pollen viability (Soares et al. 2013). Contracted PG are formed because of mutations in genes in the post-meiotic phase, which affect the androgametogenesis in Passiflora (Pagliarini 2000; Souza et al. 2003Souza et al. , 2008. The unviability of pollen can also be attributed to meiotic irregularities, such as univalent chromosomes and early and laggard chromosomes (Souza et al. 2002;Damasceno et al. 2010), as it was found here.
The culture medium used for the in vitro germination test simulates the conditions of receptivity of the stigma with the stimulus for pollen tube germination (Soares et al. 2013;Belo et al. 2015). The concentrations and the presence and absence of reagents determine the balance necessary for the success of the technique (Belo et al. 2015). Several compositions of the culture medium can be tested to arrive at the best specific protocol for each species (Vida et al. 2011). Some authors consider the addition of stimulating elements to the culture medium as a favorable factor for the germination of PG (Taylor and Hepler 1997). It has been reported in a study conducted on Actinidia chinensis var. delicious (A. Chev.) A. Chev. that when boric acid was added to the culture medium, the effect was significantly greater for the germination of the pollen in comparison to the condition in which boric acid was not added (Borghezan et al. 2011). The same result with the addition of boric acid was corroborated in a study with Passiflora species (Vida et al. 2011). Thus, the low percentage of germination of PG in the genotypes evaluated in this study may be moreover related to the late time of the collection and the environmental variables. The condition of the culture medium used in this study was also used in other studies for P. suberosa L. and P. sublanceolata (Killip) MacDougal, from which we could infer that the liquid culture medium was suitable for PG germination and also negatively pointed out the collection time hours after anthesis (Cruz et al. 2008;Belo et al. 2015).

Conclusions
The crossing of P. vitifolia vs P. hatschbachii resulted in progeny hybrid with regular meiotic behavior, high pollen viability and PG germination. The interspecific hybrids with high pollen viability, found in this study, may be useful in breeding programs as well as in the ornamental plant market because they may also be fertile and contribute with polymorphic loci in outcrossing for better use of genetic variability. The molecular tools such as SSR markers and GISH are effective to confirm the paternity of interspecific hybrids in Passiflora.