Selection of somaclonal variants of the cultivar ‘Prata-Anã’ for resistance to Fusarium oxysporum f. sp. cubense race 1

Abstract The banana tree is one of the most cultivated fruit globally; however, some diseases significantly affect its production, such as Fusarium wilt. The most appropriate measure for controlling this disease in areas with inoculum pressure is the use of resistant cultivars. Therefore, this study aimed to generate banana somaclones of the cultivar ‘Prata-Anã’ resistant to Fusarium wilt by inducing somaclonal variation. ‘Prata-Anã’ stem apexes were established in vitro in MS culture medium and, on a monthly basis, subcultivated in AIA and adenine sulfate supplemented MS medium with added plant regulators: 6-benzylaminopurine (BAP, 4 ml L-1), Thidiazuron (TDZ, 1 ml L-1), and Paclobutrazol (PBZ, 10 ml L-1). The treatments were: T0: no regulator, T1: BAP, T2: TDZ, T3: PBZ, T4: BAP + TDZ, T5: BAP + PBZ, T6: TDZ + PBZ, and T7: BAP + TDZ + PBZ. After the twelfth subculture, the regenerated plants were planted in boxes containing sterile soil infected with Fusarium oxysporum f. sp. cubense, and evaluated after 90 days for resistance to the pathogen. Somaclonal variants T2-1 and T2-2, generated in Treatment 2, with TDZ, were selected as resistant. This result is promising for the launch of a new Fusarium race 1-resistant banana variety.


Introduction
Bananas are a source of nutrients and income for thousands of families in several countries around the world, such as India, China, Indonesia, Brazil, and some countries of Africa. Brazil is the fourth largest banana producer, accounting for approximately seven million tons of fruit (FAOSTAT, 2018). Prata subgroup cultivars, such as 'Prata-Anã', and derivative selections such as 'Gorutuba' and 'Catarina', form the basis of Brazilian production, representing approximately 70% of the area in the country cultivated with this fruit (IBGE, 2016;COLTRO;KARASKI, 2019). Banana production in Brazil is characterized by small producers and a domestic consumption of 65 to 70% of total production, demonstrating its social importance in terms of food security, job creation, and subsistence (COLTRO; KARASKI, 2019).
Both the production of bananas for subsistence, and for the local and export fruit trade are threatened by the occurrence of Fusarium wilt, caused by the fungus Fusarium oxysporum f. sp. cubense (Foc) (PLOETZ, 2015;DITA et al., 2018). In addition to concerns about managing the pandemic caused by tropical race 4 (TR4) of the pathogen, some countries are searching for strategies to cope with race 1 which is endemic in the main banana producing regions of the world (ARINAITWE et al., 2019;GONÇALVEZ et al., 2019). In Brazil, there are no records of TR4; however, Foc race 1 has had a great impact, either by reducing the productivity of banana plantations or by making the soil unsuitable for planting susceptible cultivars, especially in fields under irrigation which constitute a large proportion of banana production areas (REBOUÇAS et al., 2018;GONÇALVEZ et al., 2019).
The fungus Foc is a soil-dwelling pathogen that penetrates the roots of the banana tree and colonizes the tissue, affecting xylem vessels and leading to symptoms such as progressive leaf yellowing, pseudostem cracks, leaf wilt, and plant death . In addition, the pathogen produces resistant structures called chlamydospores which enable it to survive in the soil for many years (PLOETZ, 2015). Cultural and biological disease control practices are not efficient for controlling Foc, especially with high inoculum pressure in very susceptible cultivars, but have shown promising results when associated with less susceptible cultivars with quantitative resistance . So far, the results clearly indicate that the most viable disease management strategy is the development of resistant cultivars (REBOUÇAS et al., 2018;ARINAITWE et al., 2019;GONÇALVEZ et al., 2019). However, most cultivars of the genus Musa have high levels of sterility and low seed production, which makes conventional breeding a significant challenge, despite the development of hybrids resistant to Foc race 1 by some breeding programs, such as the Honduran Agricultural Research Foundation (FHIA), and the Brazilian Agricultural Research Corporation (Embrapa) (AMORIM et al., 2011;KHAYAT et al., 2011).
The exploration of somaclonal variations that occur during the micropropagation process is an alternative strategy to progeny breeding and selection programs (CHEN et al., 2013;GHAG et al., 2014). This approach is related to the occurrence of genetic changes induced by in vitro cultivation, which can cause genetic or epigenetic variation (PENNA et al., 2019;ANIL et al., 2018). A clear demonstration of success with somaclonal variants resistant to TR4 occurred in Taiwan using the cultivar 'Giant Cavendish' (HWANG; KO, 2004). Another study, using somatic embryos kept in tissue culture for an extended period to generate somaclones which were then challenged with Foc race 1, found that four of the 26 banana cv. 'Rasthali' somaclones presented greater Foc resistance. In addition, cDNA-RAPD analysis identified at least one differentially expressed gene (lipoxygenase) in these somaclones (GHAG et al., 2014).
In the present study, 'Prata-Anã' stem apexes were subcultured in culture medium with different growth regulator combinations and successive cultivations to generate and select somaclonal variants resistant to Foc race 1.

T6
Thidiazuron TDZ (1 ml.L -1 ) + Paclobutrazol PBZ (10 ml.L -1 ) T7 6-benzilaminopurina BAP (4 ml.L -1 ) + Thidiazuron TDZ (1 ml.L -1 ) + Paclobutrazol PBZ (10 ml.L -1 ) A total of 12 subcultures were conducted at 30-to 40-day intervals. At the end of the subculture process, the sprouts were individualized and converted into plants, then transferred to MS medium without regulator for rooting. After 30 days, the rooted plants were transferred to tubes containing substrate and acclimated in a greenhouse for 30 to 45 days, until they reached a height of approximately 15 cm.

Fungus culture and inoculum preparation
This study used the CNPMF 218A isolate from Foc race 1, obtained from the biological collection of F. oxysporum f. sp. cubense from the Embrapa Laboratory of Phytopathology. This isolate was selected for its virulence and aggressive infection of banana cultivars (HADDAD et al., 2011;COSTA et al., 2015;ROCHA et al., 2020).
The Foc 1 isolate was grown on potato dextrose agar at 25 °C under a 12-hour photoperiod. After colony growth, a suspension of conidia was prepared and approximately 20 mL of this deposited in 1 kg of sterilized rice. Subsequently, the culture medium was incubated at 25 °C under a 12-hour photoperiod. After 20 days, colony-forming units (CFU) were counted using a serial dilution to verify the spore concentration and viability. The CFU were counted using a Neubauer camera. The concentration used for inoculations was 10 6 CFU g -1 of substrate (MADIGAN et al., 2016).

Preliminary selection of Fusarium wilt resistant somaclones
The experiment was conducted in a greenhouse using 310 L polyethylene boxes filled with autoclaved soil which was subsequently infected with 1 kg of the Foc inoculum. 'Prata-Anã' banana plants, approximately 15 cm high, were planted in the Foc infected soil at a density of 30 seedlings per box. Non-treated 'Prata-Anã' seedlings were planted in each box as a positive control. A total of 2,520 plants subjected to seven treatments (Table 1) were planted, with 360 plants evaluated per treatment.
After 90 days or plant death, internal symptoms of rhizome discoloration were evaluated using the grading scale (ranging from 1 to 5) proposed by Dita et al. (2014), where 1: absence of symptoms, 2: rhizome with initial discoloration, 3: rhizome discoloration throughout the vascular system, 4: rhizome with most internal tissues presenting necrosis, and 5: fully necrotic rhizome.
Plants with no symptoms of disease or with the lowest disease severity score were designated as putative Fusarium wilt race 1 resistant somaclones. These seedlings were used as the source material for further tissue culture, where they were introduced in vitro and micro-propagated for a second experiment to verify their Foc resistance.

Confirmation of resistance in selected somaclones
The micro-propagated seedlings of the somaclones supposedly resistant to Fusarium wilt (identified in the preliminary selection) were replanted in the infected soil, in a completely randomized design with ten replicates for each somaclone.
In this second experiment, the plants were evaluated for external leaf yellowing symptoms using the grading scale of Dita et al. (2014), where 1: no symptoms, 2: initial yellowing in old leaves, 3: yellowing in old leaves and initial discoloration in new leaves, 4: intense yellowing in all leaves, and 5: plant death. The symptoms were recorded from onset up to 60 days, at seven-day intervals. External leaf symptom indexes were determined from the scores and used to calculate the area under the disease progress curve (AUDPC) using the formula proposed by Shaner and Finney (1977).
Internal symptoms were analyzed using transverse rhizome sections of the plants 60 days after inoculation or after plant death, applying the same internal symptom grading scale as used in the preliminary experiment (DITA et al., 2014). The scores obtained were used to calculate internal indexes ID (%) values of Fusarium wilt symptoms.

Evaluation of fungal structures in the root tissue of the somaclones tested for Fusarium wilt resistance
Histological evaluation was performed for the putative somaclones identified in the preliminary test and for those used in the resistance confirmation experiment. In both cases, root fragments were collected after the final internal symptom evaluation during which the rhizomes were sectioned. The Phillips and Haymann (1970) protocol was used for root staining. The root fragments were soaked for 1 h in a 10% potassium hydroxide (KOH) solution in a water bath at 90 °C. Thereafter, the solution was discarded and the biological material washed in water to remove all the KOH. The roots were then transferred to a 1% HCl solution for a 5 min period. The HCl solution was discarded and the roots were stained using Trypan Blue in a 0.05% lactoglycerol solution (2:1:1, lactic acid: glycerin: water) and boiled in a microwave for 40 sec after which the dye was discarded. The roots were then immersed in lactoglycerol solution (2:1:1; lactic acid: glycerin: water) to remove excess dye. The root fragments were visualized and microphotographed under a light microscope.

Statistical analysis
Regarding the preliminary test internal symptom evaluation scores, the frequency of each score per plant was calculated and converted to a percentage. To study the treatments, a cluster analysis using color was also performed, based on a heat map prepared by transforming the internal symptom scores into a binary matrix, where 0 was assigned to plants without symptoms and 1 to plants with disease symptoms. The data for each plant was represented by a color, where green was associated with resistance and red with susceptibility. The ID symptoms and AUDPC data obtained from the second experiment were subjected to analysis of variance, and the means were grouped by the Scott-Knott test at 5% significance. The analyses were performed using the R software.
Rev. Bras. Frutic., Jaboticabal, 2020, v. 42, n. 6: (e-620) Owing to contamination problems and losses during the in vitro micro-propagation process for the second experiment, only four plants were regenerated and used in the next stage. These plants were designated as putative Fusarium wilt race 1 resistant somaclones and named T2-1, T2-2, T4-1, and T5-1. Contamination during the micropropagation process may be related to Foc structures in the tissues of plants which, although not manifesting symptoms, may have been infected. Some resistance mechanisms are known to form after infection such that the pathogen enters the tissues, but cannot establish itself because the plant resistance response is faster (LI et al., 2015). The presence of toxins released by the fungus in rhizome tissues is also likely, since Foc is a necrotrophic fungus that releases toxins to destroy host tissues (PLOETZ, 2015). Some of the main phytotoxins produced by Foc are fusaric acid and beauvericin (PORTAL et al., 2018).
A low number of somaclones without symptoms was generated in this study. This result was expected as genetic variation induction is known to occur in low percentages. A similar study in Taiwan induced somaclonal variation in 20,000 Cavendish banana seedlings planted in a field infected with Foc TR4 and reported approximately 0.01% of genotypes resistant to Fusarium wilt. This particular study registered GCTCV-218, a somaclonal variant of Cavendish resistant to Fusarium wilt, for commercial cultivation under the name Formosona, and two other Cavendish somaclonal variants named GCTCV-53 and GCTCV-119 were identified (HWANG; KO, 2004). The findings of the present study indicate that the use of plant regulators in association with successive in vitro multiplication may lead to greater efficiency in generating resistant somaclonal variants because the study conducted in Taiwan showed a lower percentage of resistant variants (0.01%) than were found in the present study (0.079%).
A study by Ghag et al. (2014) regenerated four genotypes of cv. 'Rasthali' banana plants resistant to Fusarium wilt using the somaclonal variation tool. In this study, embryos were repeatedly subcultured and maintained for 14 years under tissue culture conditions. This indicates a low frequency of mutations acting on resistance, requiring a significant number of plants or a long period of subcultures to increase the chances of selecting possible somaclonal variants that are resistant to the disease. Figure 2 shows the incidence of plants with symptoms (highlighted in red) and plants not showing symptoms (highlighted in green) and indicates that a small amount of genetic variability occurred for the resistance factor in the 2,520 somaclonal variants evaluated, considering that most of them showed Foc susceptibility ( Figure 2). The treatment with the highest number of plants without symptoms was T2 (TDZ), represented by the T2-1 and T2-2 somaclones. This finding indicates that TDZ used on its own and not combined with other plant regulators in the culture medium was most effective in the generation of somaclonal variants resistant to Foc race 1. Thidiazuron is a plant regulator of the cytokinin group, responsible for cell division and used in many studies to induce organogenesis and embryogenesis (PELAH et al., 2002;SHEIBANI et al., 2006;AHMAD et al., 2018).  A study by Goelzer et al. (2019) showed that a combination of TDZ and ANA had a negative effect on sprout growth, suggesting that the use of TDZ alone in the culture medium intensified cytokinin effects on cell division, which could cause DNA polymerase errors, consequently generating mutations. This may account for why, in this study, T2 (TDZ) emerged as the best treatment for generating possible genetic variation in comparison with the other treatments that used combinations of plant regulators.
Analysis of the second experiment results confirmed the resistance of the somaclonal variants T2-1 and T2-2. Evaluation of the external symptoms of somaclones T2-1 and T2-2 by AUDPC yielded a value of 250. This was statistically different from the AUDPC values of somaclones T4-1, T5-1, and T0 (control), which had AUDPC values ranging from 600 to 1,500 ( Figure  3). As for the internal symptoms evaluated by the ID, the somaclonal variants T2-1 and T2-2 had the lowest indexes (0% and 5%, respectively). These values were statistically different to the ID values of somaclones T4-1, T5-1, and T0, which ranged from 40 to 70% ( Figure  3). The transverse sections of plant rhizomes showed characteristic Fusarium wilt symptoms, with reddish brown coloration in the vascular tissues in T4-1, T5-1, and T0 (Figure 3 A, D, E). Through these analyses, somaclones T4-1 and T5-1, which appeared to be Foc resistant during the preliminary test, were shown to be susceptible to Fusarium wilt during the second experiment.
The histological evaluation at the end of the preliminary test revealed Foc structures such as chlamydospores and hyphae in the tissues of T0 and somaclonal variant T2-2 only (Figure 4 A and C). At the end of the second experiment to confirm resistance, the histological evaluation showed that somaclonal variants T2-2, T4-1, T5-1 and T0 had Foc structures inside the root tissues. However, the T2-1 variant lacked evidence of fungal structures in the tissues, corroborating the preliminary test result for this somaclone (Figure 4 B and G). Considering that the absence of pathogen structures in root tissues indicates that the pathogen has not successfully penetrated these tissues, the resistant variant T2-1 may have developed physical and chemical barriers to block pathogen penetration (PETIT-HOUDENOT et al., 2017;BANI et al., 2018). As reported in a histological study by Costa (2013), the cultivar 'BRS Platina', a Foc race 1 resistant 'Prata' cultivar (launched by Embrapa in 2011), showed no Foc structures inside the roots, a finding similar to that for somaclonal T2-1. The T2-2 somaclonal variant showed no external or internal Fusarium wilt symptoms, indicating Foc resistance; however, it had pathogen structures in the root tissues. These findings suggest that, whereas this somaclone has no resistance to Foc tissue penetration, it has resistance mechanisms that prevent pathogen infection, possibly because the genetic resistance of plants corresponds not only to the capacity to prevent, but also to delay disease development (VALE et al., 2001;LI et al., 2015).
The leaves of the somaclonal variants showed morphological differences. Somaclones T4-1 and T5-1 had reddish spots on their leaves, similar to the spots on the T0 leaves ( Figure 5). The leaves of somaclones T2-1 and T2-2 had no such reddish spots, suggesting some form of genetic modification in these two variants. Previous studies have reported morphological changes in banana trees in relation to somaclonal variation (ALVARES et al., 2002). Rev. Bras. Frutic., Jaboticabal, 2020, v. 42, n. 6: (e-620) The pathogen Foc race 1 is present in most banana fields in Brazil. The 'Prata-Anã' cultivar is susceptible to this pathogen and the disease it causes has led to a 60% loss in production in plantations around the country. This fungus produces chlamydospores, resistant structures that can survive long periods in the soil (PLOETZ, 2015;PEGG et al., 2019). Cultural, chemical, and biological control methods have not been effective in controlling the disease (TUSHE-MEREIRWE et al., 2000;KARANGWA et al., 2016), consequently, identification of a 'Prata-Anã' somaclone resistant to Fusarium wilt race 1 that maintains other agronomic and sensory characteristics of the original cultivar is essential to protect the cultivar and ensure that it continues to be a source of income for farmers, especially given its preference by Brazilian consumers (IBGE, 2016;MOSTERT et al., 2017;HADDAD et al., 2018).
The findings of this study demonstrate that somaclones T2-1 and T2-2 are promising variants for further studies by the Embrapa banana genetic improvement program. Such studies should include both agronomic characterization to confirm the permanence of desirable characteristics similar to those of the 'Prata-Anã' cultivar, and the elucidation of resistance mechanisms generated in these somaclones. The results of current and future research on the somaclones generated in this study will have a significant impact on the launch of a new Fusarium race 1 wilt resistant cultivar.

Conclusion
Induction of somaclonal variations of the 'Prata-Anã' cultivar is viable. as indicated by the generation of the somaclones T2-1 and T2-2, which showed evidence of resistance to Fusarium wilt race 1.