Bananas and banana trees (Musa spp.) are among the most important crops in the countries of the tropics and subtropics. Worldwide, they are the fourth most important, after rice, wheat and maize (^{FAO, 2017}). However, their production is threatened by diseases such as the Panama disease or the Fusarium wilt. This disease, caused by Fusarium oxysporum f. sp. cubense (FOC), is one of the most destructive and economically important in the genus Musa (^{Ploetz, 2015}). The control of this disease with agrochemicals is costly and causes serious damages to the environment. Due to this, in recent years, breeding for resistance has been considered the most adequate solution. Every breeding technique requires a procedure for the positive selection of genotypes with the best traits (^{Buddenhagen, 2009}). In earlier investigations, a method was developed to differentiate the resistance of banana tree cultivars at FOC race 1, at a foliar level, with the use of filtrates of the fungus culture (^{Companioni et al., 2003}). However, the results for their introduction into the breeding programs in the crop on a massive scale need to be validated for different races of the fungus (isolations belonging to the group of plant compatibility (VCG) [01210] race 1; and the isolation belonging to VCG [0124/125] race 2. The present work was carried out with the aim of evaluating the method for the rapid differentiation of the Panama disease using the filtrate of the FOC culture FOC race 1 VCG [01210] and race 2 VCG [0124/125].

Research was carried out in the Plant-Pathogen Interaction Laboratory of the Ciego de Ávila “Máximo Gómez Báez” University, Cuba.

Plant material. We used foliar samples of banana trees eight to nine months after being planted in the Banana Germplasm Bank of the National Institute of Research in Tropical Foods of Santo Domingo, Villa Clara, Cuba. All banana cultivars used were previously classified as resistant or susceptible to FOC race 1 and 2, by ^{Pérez et al. (2004}).

Fungal strains. FOC race 1 VCG [01210]; and race 2 VCG [0124/125], from the strain collection of the Plant Health Research Institute (INISAV), Havana, Cuba.

Statistical analysis. The statistical processing of data required the use of the Statistical Package for Social Sciences (SPSS for Windows, version 8, Copyright SPSS Inc., 1989-1997).

In the first part of the investigation, fungus culture filtrates were obtained during the in vitro growth of FOC race 1 VCG [01210]; and race 2 VCG [0124/125]. For this purpose, the optimum moment for the harvest of the fungus culture filtrate of races 1 and 2 was determined on a daily basis for 30 days. The following determinations were carried out: absorbance of the fungus culture filtrate (A), fresh mass of the fungus culture (g) and the phytotoxic activity of the fungus culture filtrates.

Filtrations of the fungus culture (FC). In order to obtain the culture in a liquid medium of the FOC strains corresponding to races 1 and 2 of the fungus, we took mycelium discs, 8 mm in diameter near the periphery of the culture grown in Petri dishes with a Potato Dextrose Agar culture medium. We added 100 mL of Czapek culture medium, modified according to ^{Companioni et al. (2004}) with a pH of 5.5, in 250 mL Erlenmeyer flasks. Each Erlenmeyer was inoculated with a mycelium disk and incubated at 28±2 °C, 56 µmol.m^-2.s^-1 of light density and with a photoperiod of 12 hours light/12 hours darkness. The fungal FCs were harvested on a daily basis for 30 days for both fungal races. In each one of the evaluation times, three Erlenmeyer flasks were collected with a volume of 100 mL of inoculated culture medium. The liquid culture medium was filtered in each of the times evaluated through four layers of fat, and then through Whatman No. 1 filter paper. The fresh mass of the fungal culture was then determined in order to determine the growth of the microorganism. Before determining the absorbance and the phytotoxic activity of the fungus FC, the FC was centrifuged at 8000 rpm for 20 minutes, using a Sigma-201M centrifuge to eliminate all the remaining mycelia and conidia. The supernatant was poured through filters with pores, 0.22 mm in diameter (Sartorius, NG, Göttingen, Germany).

Absorbance of the fungal FC. The fungal FCs obtained for both pathogen races were analyzed for absorbance in the UV region of the spectrum. For this, 1 mL aliquots of the fungal FC were diluted 1/25 in sterile distilled water, and the absorbance was measured from 250 to 300 nm in a UV-Vis spectrophotometer (LKB-Pharmacia).

Phytotoxic activity of the fungal FC. This was determined with the bioassay on banana detached leaves according to ^{Companioni et al. (2003}) in each time evaluated (1 to 30 days). Mid-aged leaves were collected, according to their position on the plant (leaves 3, 4, 5) of the cultivar Gros Michel (group AAA, susceptible to race 1); and Bluggoe (group ABB, susceptible to race 2). They were washed for 20 minutes with a commercial detergent, rinsed with sterile distilled water, and dried using Whatman No. 1 filter paper. Holes were made with a sterilized needle on the adaxial side of the limbo of the leaf with a 3 cm separation between then. After the damage, 5 µL of the fungal FC of the fungus gathered in each time evaluated, were added. The leaves were incubated at 28±2 °C, 56 µmol.m^-2.s^-1 of light intensity, a photoperiod of 12 hours light/12 hours darkness and a relative humidity of 70% for 48 hours. After this time, the phytotoxic activity of the fungal FCs was evaluated, through the symptomatological expression of necrosis formed around the point of application of the fungal FC, expressed as the area of the elliptical lesion (mm²).

The Czapek medium, not inoculated and incubated as described, was used as a control treatment for all the experiments evaluated in this work. In this experiment, each treatment included three leaves from different plants (18 lesions/leaf). Later, we determined the response of Musa spp. Cultivars to the application of the fungal FC of races 1 and 2, evaluated by the bioassay on banana detached leaves following ^{Companioni et al. (2003}). Fifteen banana cultivars with different levels of resistance to the in vivo pathogen were compared, following ^{Pérez et al. (2004}). The FC of race 1 of FOC was harvested after 15 days; and race 2, 29 days after inoculation in the Czapek medium.

Based on the absorbance values (250-300 nm) of the FC of the fungus obtained during the in vitro growth of the strain of FOC race 1, there are three crucial moments of excretion of metabolites into the culture medium, which correspond to days 15-16, 21-23 and 29-30 days of cultivation respectively (Figure 1). The highest level of absorbance was registered 15 days after planting, with 0.340. However, in the FOC race 2 culture filtrates, the highest absorbance levels were found on days 29 and 30 of cultivation (0.755 and 0.724 respectively), which resulted higher than those reached by FOC race 1 on day 15 of cultivation (Figure 2). The species of the genus Fusarium produce a wide variety of secondary metabolites (^{Sieber et al., 2014}). It is also well-known that the expression of the microbial metabolism is not necessarily related to the maximum increase of the myceliar mass, but rather to different moments in the growth of the microorganisms, depending on its sources of nourishment and the conditions of its in vitro cultivation (^{Etzerodt et al., 2016}).

The cultivation conditions for both FOC strains helped the excretion into the medium of extracellular metabolites produced by the pathogen in different moments of its growth in vitro, which could be related to the pathogenicity and/or the virulence of these pathogenic strains.

When the FOC race 1 [VCG 01210] was cultivated in a liquid medium, no rapid increase in the growth of the evaluated microorganism was found on the fresh mass for the first 18. This result may be related to a slow adaptation of the fungal strain to the culture medium, as well as to other conditions established for their growth and development. After day 18, the pathogen grows exponentially, and a maximum of 12.2 g was recorded for its fresh mass 22 days after inoculation (Figure 3A). However, related to the excretion of extracellular metabolites by the pathogen, the effect of the FC of the fungus on the responses of the foliar tissues of susceptible plants displayed notorious differences in diverse fungal cultivation times (Figure 3C). In general terms, although several moments can be observed with light phytotoxic activity of the FC, the highest levels were found on cultivation days 15-16, 21-23 and 29-30. The highest levels of phytotoxic activity match the increases in the absorbance of the culture filtrate and therefore, the synthesis of extracellular microbial metabolites. In general terms, the FC obtained showed specificity in their way of action. The highest levels of phytotoxicity evaluated, such as the area of the lesion on the susceptible genotype, were obtained when the FC cultivated on days 15-16 during the growth of FOC race 1 in the Czapek medium. According to ^{Brakhage (2013}), the production of microbial metabolites takes place in different phases of the growth curve of microorganisms, broadly linked to a particular phase for each one of them, yet highly influenced by the conditions of cultivation. On the other hand, ^{Zhang et al. (2015}) showed that the microbial phytotoxins are metabolic products of the pathogen, characterized as pathogenicity and/or virulence factors. Based on the results obtained in the present investigation, these metabolical products are synthesized by the pathogen in a certain growth phase. The growth curve of the race 2 fungus displays increases in the fresh mass of the microorganism starting on day 9 after cultivation, reaching maximum values 18 days after inoculating in the culture medium (16.1 g) (Figure 3B). This displays a greater adaptation and assimilation of the components of the culture medium under the conditions established for its growth in vitro in relation to the strain of FOC race 1. Meanwhile, FOC race 2 is able to constantly produce phytotoxic molecules 10 days after it begins growing under the same conditions established for both strains (Figure 3D). However, although the evaluation of the phytotoxic activity shows the constant excretion of phytotoxic metabolites into the culture medium, there may be a process of conversion of the initially synthesized components into others that still maintain the displayed phytotoxic activity. The Bluggoe cultivar (susceptible) displays damages, measured as area of the lesion, in a continuous manner with the FCs of 11 to 30 days of cultivation of the fungus. Nevertheless, 29 days after the cultivation of the pathogen in the liquid medium, a greater damage was observed on the tissues of susceptible plants. Several researchers, on multiple plant-pathogen interactions, have described the extraction of culture filtrates with phytotoxic activities for their use as precocious selection agents for resistance. In addition, schemes have been developed for the isolation and purification of the metabolites involved in the selective response of plants (^{Ramírez et al., 2015}). The precocious selection of resistance in plants faced with different ‘formae specialis’ of Fusarium oxysporum has been a crucial objective in conventional and biotechnological breeding. Raw extracts and pure metabolites of FOC cultivations have been used in the selection of resistant plants (^{Companioni et al., 2003}; ^{Saraswathil et al., 2016}).

Figure 1. Excretion to the metabolite culture medium during the growth of Fusarium oxysporum f. sp. cubense race 1 VCG [01210] in a modified Czapek medium. 

Figure 2 Excretion to the metabolite culture medium during the growth of Fusarium oxysporum f. sp. cubense race 2 VCG [0124/125] in a modified Czapek medium. 

Figure 3. Phytotoxic activity of filtrates of crops planted on a daily basis during the in vitro growth of Fusarium oxysporum f. sp. cubense races 1 and 2. Fresh mass of the pathogen (A and B), and area of the lesion in banana tree cultivars susceptible to the disease in vivo (C and D). 

The cultivars observed to be the most susceptible to the FC of race 1 were Manzano Criollo, Gros Michel, Pisang Lilin and Yangambi km 5, with no statistical differences between them, followed by the del cultivar Paka. All FHIA cultivars evaluated (01, 02, 03, 04, 18 and 21) were observed to be resistant to the phytotoxic action of the FC race 1 VCG [01210], without statistical differences in their respective responses. Cultivars Burro Criollo, Pelipita, Pisang jari guaya and Bluggoe also proved to be resistant, with a similar response to those of the FHIA (Table 1).

^{Pérez et al. (2004}) determined, among other studies, the reaction of different cultivars towards the FOC populations in Cuba, and the frequency of diseased plants. Their results show that race 1 most frequently affects the cultivars Gros Michel, Yangambi km 5, Paka and Manzano Criollo, while Pisang Lilin, Pisang jari buaya, FHIA03 and FHIA-18 showed the least effects. In the results described in this work, the bioassay of the FC of the race 1 fungus on Pisang Lilin cultivar leaves displayed a reaction of susceptibility, as opposed to the response of tolerance, determined previously under field conditions by Pérez et al. (2004). On the other hand, the 29-day FCs of FOC race 2 caused the greatest damage on cultivars Burro Criollo, Pisang jari guaya, FHIA-03 and Bluggoe, while in cultivars Manzano Criollo, Gross Michel, Pelipita, Pisang Lilin, Paka and Yangambi Km 5, they displayed a similar resistance response to the evaluated FHIA. The bioassay of the cultivar Pisang jari buaya revealed a susceptibility response, unlike the tolerance displayed in the mentioned field evaluations. In this sense, the problem concerning the resistance tests under in vivo is the task of ensuring a homogenous distribution of the inoculant of the pathogen in all plants to be evaluated. Some plants may have received an excessive amount of the inoculant, while the others may not receive enough inoculant to develop the disease (^{Mert and Karakaya, 2003}). Additionally, there are concerns over the effect of the environment on inoculated plants studied in the field (^{Ribeiro et al., 2011}). On the other hand, it is not always possible to characterize the FOC populations using pathogenicity tests, due to the interactions of the plant with the environment. There is proven evidence of the differential response of plants under different environmental conditions to established differential groups (^{Li et al., 2013}). The FCs of 15 and 29 days obtained from FOC strains of races 1 and 2, were able to differentiate over 93% of the individuals with a known response to the pathogen in vivo. The remaining 7% consisted, in both cases, of an intermediate response to the disease under natural conditions. It is important to point out that, in order to establish a method to evaluate the resistance to diseases, it must be better than the traditional one in terms of work, space and time, aspects which are considered in this method, tested for both races of the pathogen.

The results obtained in this work show that the method to differentiate susceptibility or resistance to the Panama disease according to ^{Companioni et al. (2003}) can be performed in a quick and non-destructive way with the use of the FC of the fungus treated in on leaves for different populations of the pathogen, which may be applicable to the breeding programs of Musaceae, both conventional and biotechnological. On the other hand, they are the first set of results focused on the validation of the method that break the natural biological scheme of the natural cycle of the disease. In addition, they lay the foundations for future projects related with the identification of phytotoxic metabolites, and of the specificity of alleged avirulence genes of the pathogen, which will help determine the role played by the metabolites involved in the response of Musa spp. to wilt by Fusarium.