Experimental design

The experimental design was carried out taking the following into consideration:

1) Each finca, or property, was taken as an independent pathogenic population, the individuals of which were represented by the multiple spores or individuals produced in the infectious cycle of the disease. 2) If the phenotype is considered the expression of the genotype based on a given environment, and that in M. fijiensis´ environment, the pressure exerted by the fungicides favors the genotypes that help tolerate, to a greater or lesser degree, the action of these fungicides, it would be therefore expected that if there is a favored genotype, it would express itself through phenotypical traits, identifiable among individuals. 3) Previous evaluations have shown that different phenotypical traits repeat themselves between individuals, indicating a probable relationship between these traits. 4) The active ingredients evaluated have been used commerically on the field, therefore the population and the pathogenic individuals of this study have already undergone the selection pressure of these fungicides. Based on these conditions, in vitro experiments were designed to compare if the population sensitivity (property) or the individual sensitivity (monosporic) to piraclostrobin, epoxiconazol, and tridemorf, is related to the phenotypic behavior of each individual.

Phenotypic Variables Evaluated

A phenotypic characterization was performed which helped divide the isolates into two groups, one made up of quantitative variables, and the other, by six qualitative variables (Figure 1). The first group was composed of the phenotypic data, qualified quantitatively: length of the germ tube and biomass production (weight of the mycelium); the second group was composed of qualitative variables: 1) number of germ tubes (one germ tube only, two germ tubes of similar lengths or two unequal germ tubes in which one is less than half as long as the other. 2) Branching of the germ tube (germ tubes, branched or not branched. 3) Color of the mycelium (white, pink, or gray), 4) Morphology of the mycelium (circular division on the mycelium, parallel division on the mycelium, irregular division and indefinite division, 5) Presence of exudate (exudate present or absent, 6) Shape of the mycelium (ovoidal or undefined "amorphous").

Figure 1 Qualitative phenotypic variables. From left to right: A) White, undefined morphology, with exudate, amorphous shape. B) Gray, parallel morphology, no exudate, oval-shaped. C) white, undefined morphology, no exudate, amorphous shape. D) white, parallel morphology, with exudate, oval-shaped. E) gray, circular morphology, no exudate, oval-shaped. F) pink, irregular morphology, with exudate, oval-shaped. G) pink, circular morphology, with exudate, oval-shaped. H) pink, parallel morphology, with exudate, oval-shaped. I) gray, undefined morphology, with exudate, oval-shaped. J) two germ tubes of similar length with ramifications. K) one germ tube, without ramifications. L) two germ tubes of similar length where one is less than half the length of the other, without ramifications. 

Plant Material

Plant samples were used with symptoms of being in stage six of the disease (final stage of the disease, characterized by sexual recombination and the production of spores), obtained during 2013 and 2014, from 12 banana-producing properties in Colombia, seven of which are found in the Urabá region, and six in the Santa Marta region. Out of the 12 properties, 10 were commercial properties, and therefore the pathogenic population found there has been constantly placed under the pressure of fungicides, and they are expected to have lost, to a greater or lesser degree, their sensitivity, whereas in the two remaining properties (named as wild fincas) are not commercial properties, so the pathogenic populations found there have not undergone the pressure of fungicides and show high levels of sensitivity towards the different active ingredients.

Determining the Sensitivity of M. fijiensis Populations to Fungicides

Initially, the sensitivity was measured in each property using the protocol endorsed by FRAC (Fungicide Resistance Action Commitee) (^{FRAC, 2008}) for the evaluation of systemic fungicides in M. fijiensis, called "Ascospore germ tube elongation," based on the measurement of the length of the germ tube, which helps value the level of population sensitivity from the average of individual measurements. Both the concentrations and the number of individuals or spores measured were fixed according to FRAC specifications. For fungicides QoI to which piraclostrobin belongs, three concentrations were evaluated for every property (0-1-10 mg/L) and 600 spores were measured for each property. For the two remaining active ingredients, the specifications given for SBI (Sterol Biosynthesis Inhibitors)-type synthesis were taken, to which both epoxiconazol and tridemorf belong. In this way, for epoxiconazol, for every finca 210 spores were evaluated in five concentrations (0-0.01-0.03-0.1 and 1 mg/L). For tridemorf, six concentrations were evaluated (0-0.1-0.3-1-3 and 10 mg/L) and 240 spores for every finca.

Monosporic Isolates

After establishing population sensitivity and from this setup, germinated spores were selected at random in the absence of the fungicide (0 mg/L), which were previously quantified for length, the existence of ramifications, and number of germ tubes. These spores were isolated using a light microscope and a glass capillary tube, and transferred to a PDA nutritional medium where they grew at room temperature (26±1 °C) for 40 days until a monospore culture was obtained. After this time, the color, morphology, the presence of an exudate, and the shape of the mycelia were determined. Next, and taking advantage of the compact and non-extended growth of M. fijiensis under in vitro conditions, the total mycelia were weighed using an analytic scale, and the biomass produced was quantified. A total of 123 monospore cultures was obtained, and comparisons were carried out on these.

Determining the Individual Sensitivity of M. fijiensis to Fungicides

Out of the 123 monospore cultures, those with the greatest differences in the length of the germ tube and weight of the mycelia were chosen (since the sensibility measurements were based on these traits) and a total of 41 monospore isolates was obtained; their individual sensitivity to the three active ingredients was quantified using spectrophotometric microtechnics (Table 1). This technique is also endorsed by FRAC, but not for M. fijiensis, since the absence of sexual sporulation of this pathogen in vitro implies that this evaluation must be performed on monosopre cultures or on a limited number of individuals that would not represent the population. In this way, this technique would not be useful for determining the sensitivity of a population, but it would be to determine the sensitivity of an individual from that population. This technique was based on the preparation, based on the mycelia of each one of the 41 monospore isolates, a suspension at a concentration of 10⁴ mycelium fragments/ml, which was incubated for eight days in eight ELISA dishes with 96 pits. For each active ingredient, four concentrations were evaluated: piraclostrobin (0-0.1-1-20 mg/L); epoxiconazol (0-0.1-1 and 10 mg/L) and tridemorf (0-1-10 y 100 mg/L). These concentrations were fixed from previous evaluations in which the inhibition range was established for each active ingredient when this technique was used.

Table 1 Phenotypic traits for the 41 monospore isolates (31 from commercial properties and 10 from wild properties) which were measured for sensitivity via measurement of the germ tube and microtechnics. (A) Isolation. (T) Number of germ tubes. (R) Ramification. (C) Color (bl:white. ro: pink. gr: gray). (M) Morphology (ci: circular. pa: parallel. in: undefined. irr: irregular). (E) Exudate (au: absent. pr: present). (F) Shape(am: amorphous. ov: oval). (um) Length germ tube. (mg) Biomass weight. (P.t) IC₅₀ piraclostrobin germ tube. (E.t) IC₅₀ epoxiconazol germ tube. (T.t) IC₅₀ tridemorf germ tube. (P.m) IC₅₀ piraclostrobin microtechnics. (E.m) IC₅₀ epoxiconazol microtechnics. (T.m) IC₅₀ tridemorf microtechnics. 

Quantification of Sensitivity

The parameter used to quantify the sensitivity, both at a population and an individual level, was the IC₅₀ "half maximal inhibitory concentration", calculated from a linear regression of two points (Y = mX + b), where the points graphed were the concentrations above and below, between which would lie 50 % of inhibition and the resulting curve for each fungicide was represented in an Excel spreadsheet, along with the IC₅₀ estimated. The estimated IC₅₀ was the dose or concentration (X axis) produced by 50 % inhibition (Y axis) in the curve (^{Chinn, 1997}).

Statistical Analysis

In order to establish if there were significant differences between the expression of the different phenotypical features, as well as between the expression of these ranges with individual or population sensitivity to the fungicides evaluated, variance analyses were carried out with 95 % confidence to compare between the factors that represent the quantitative variables: a) Length of the germ tube, b) Production of biomass, c) Population or individual sensitivity to each active ingredient (IC₅₀). And the factors that represent the qualitative variables: a) Number of germ tubes, b) Branching of germ tube, c) Color of mycelium, d) Morphology of the mycelium, e) Presence of exudate, f) Shape of mycelium. Accepting the statistical significance as long as the P-value is less than 0.05 (Table 2). Likewise, multiple comparison tests within the factors were carried out using Fisher's minimum significant difference method (LSD) to define if there were significant differences between the averages obtained for each quantitative variable when compared to the presence of each trait that the qualitative variables represent (Table 3).

Tabla 2 P-value obtained from the variance analysis that compared quantitative variables (T.g: length germ tube; P.b: biomass production; In: individual sensitivity to fungicide; Po: population sensitivity to fungicide) and qualitative variables (y). And when comparing only between quantitative variables (z). From the moment the P-value was lower than 0,05 was there a statistical significance. 

Table 3 The traits evaluated are shown, along with the number of individuals with each feature, its population percentage, the statistical average of the germ tube length in micrometers (u), the statistical average of the weight of the biomass in milligrams (v), the statistical average of the IC₅₀ for the phenotypical traits related to the sensitivity of the population to piraclostrobin (w), epoxiconazol (x), tridemorf (z), and the individual sensitivity to tridemorf (y). (*) indicates significant difference when comparing factors. 

Active Ingredients

The active ingredients evaluated (piraclostrobin, epoxiconazol, and tridemorf), were taken from products belonging to the company BASF Química S.A. and used in the banana industry under the commercial names of COMET (Piraclostrobin 250 g/L), OPUS (Epoxiconazol 75 g/L), CALIXIN (Tridemorf 750 g/L).

Phenotypic Traits related to the Length of the Germ Tube

The variance analysis showed differences between factors for the number of germ tubes (Table 2y). In this way, within factors, when comparing averages, it was observed that spores that generated two germ tubes, but in which one reached less than half of the other's length, displayed a significant relation with germ tubes of greater length (X=216 um). However, when two tubes were presented with similar sizes, they related to a shorter length (X=150 um). Spores that only generated one germ tube were related to intermediate lengths and showed no significant differences with those spores with two germ tubes (X=185 um). The individuals that presented two similar-sized germ tubes, and therefore the shortest length, were the majority of the population, and represented 55 % of it, whereas those with greater lengths represented 34 % of the population (Table 3u). The existence of ramifications was another factor that presented significant differences between factors in relation to the length of the germ tube (Table 2y). Within the factors, and when comparing the averages, the appearance of these ramifications was significantly more related to longer germ tubes (X=208 um) than to shorter tubes (X=159 um). Individuals that presented ramifications and a greater length were the majority of the population, and accounted for 59 % (Table 3u). Germ tube length also displayed significant differences between factors with the mycelium color (Table 2y). Within the factors, and in the comparison of averages, it was observed that a greater length gave rise to a white mycelium (X=210 um), while a shorter length generated pink mycelia (X=160 um). No significant differences were observed when the mycelium obtained was gray (X=180 um). The gray mycelia related to medium-sized germ tube spores were the majority in the population, with 53 %. Meanwhile, the white mycelia related to a greater length accounted for 24 % of the population, a similar percentage to the 23 % of the pink mycelia related to a shorter length (Table 3u).

Phenotypic Traits Related to Biomass Production

There were significant differences between factors for the amount of biomass produced by each individual and the color of the mycelia (Table 2y). Within the factors, and in the comparison of the averages, individuals that weighed less corresponded to those colored gray (X=139 mg), white ones were related to a heavier weight than gray (X=173 mg), although lighter than pink (X=210 mg), which was the color related to a greater production of biomass. Proportional distribution indicated that 53 % of all individuals showed gray mycelia, and therefore lower weight, while 24 % presented white mycelia, and the remaining 23 % displayed pink mycelia and a higher weight (Table 3v).

Phenotypic Traits Related to Sensitivity to Fungicides

The level of sensibility to the evaluated fungicides was established using an IC₅₀ found by spectrophotometric microtechnics from monospore crops that helped find a response for individuals, as well as an IC₅₀ found from the average of the lengths reached by the germ tubes of a particular number of spores that helped find a response for the population.

Piraclostrobin

Individual Response. Among the factors, no significant differences were found between the sensitivity of an individual to piraclostrobin and the phenotypical characteristics evaluated (Table 2y).

Population Response: Among the factors there were significant differences for the level of population sensitivity to piraclostrobin and the number of germ tubes (Table 2y). Within the factos, and in the comparison of averages, spores with only one germ tube were found to be significantly more related to a population with a lower sensitivity to piraclostrobin (IC₅₀=0,805) than spores with two germ tubes, the latter related to a more sensitive population, regardless of one of the tubes being less than half as long as the other (IC₅₀=0,697). Proportional distribution indicated that 89 % of the population presented two germ tubes, and therefore a lower population IC₅₀, whereas the remaining 11 % presented only one germ tube and a higher population IC₅₀ (Table 3w). Also among the factors, significant differences were found with the production of exudate (Table 2y) and within the factors, comparing factors proved that isolated spores of a population with a lower IC₅₀ (IC₅₀= 0,699), and therefore a greater sensitivity to piraclostrobin, were significantly more related with the production of exudate in their mycelia, while the increase in the IC₅₀ (IC₅₀= 0,767) was related with no production of exudate. Proportional distribution indicated that 59 % of the population presented exudates, whereas the remaining 41 % did not (Table 3w). Among factors, the shape of the mycelium of the monosporic crop also displayed significant with the sensitivity of the population to piraclostrobin (Table 2y), and within the factors it was found, when comparing averages, that the spores from a population with a lower IC₅₀ (IC₅₀= 0,703) were significantly related with an oval shape in their mycelia, whereas if the population displayed a greater IC₅₀ (IC₅₀= 0,763) the remaining mycelium was more related to a lack of any particular shape (amorphous). Proportional distribution indicated that 72 % of the individuals presented an oval shape, and were therefore related to a greater sensitivity to piraclostrobin, while the remaining 28 % displayed no particular shape (Table 3w).

Epoxiconazol

Individual Response. Among the factors, no significant differences were found between the phenotypic characteristics evaluated and the sensitivity of the monospore individuals to epoxiconazol (Table 2y).

Population Response. Among the factors no significant differences were found (Table 2y), although within the factors and comparing averages, it was observed that population sensitivity to epoxiconazol displayed significant differences for the trait exudate production, since isolates from a more sensitive population to this active ingredient (IC₅₀=0,18) displayed exudate, while less sensitive isolates did not (IC₅₀=0,23). Proportional distribution indicated that 59 % of the population presented exudate, and therefore a higher sensitivity to epoxiconazol, while the remaining 41 % displayed no exudate (Table 3x).

Tridemorf

Individual response. Among the factors, significant differences were found for the existence of ramifications in the germ tube(s) (Table 2y). And within the factors, when comparing averages, it was observed that the presence of ramifications was significantly related to a lower sensitivity to tridemorf (IC₅₀=32), while the absence of ramifications was related to a higher sensitivity (IC₅₀=11,5). When analyzing the proportional distribution, it was found that individuals without ramifications were only a few in the population (41 %), while the individuals with ramifications and a lower sensitivity were the majority (59 %) (Table 3y). Among the factors, significant differences were also observed for the color of mycelia (Table 2y), and within the factors, when comparing averages, a lower sensitivity of the monospore culture to tridemorf (IC₅₀=37,5) was noticed to be significantly related to a gray mycelium, whereas if its color is white (IC₅₀=13) or pink (IC₅₀=14,8) it relates to a higher sensitivity. When the proportional distribution was evaluated, grey isolates related to a lower sensitivity to tridemorph were found to be the majority and accounted for 56% of the population, while white and pink isolates were found at equal proportions of 22 % (Table 3y). Within the factors, when comparing the averages for the trait of morphology, it was noticed that if a monospore culture presented a mycelium with an irregular morphology, this individual would be significantly more related to a higher sensitivity (IC₅₀=8,17), since in vitro it would require a lower concentration of tridemorf to achieve the same level of inhibition than a monospore culture with an undefined morphology (IC₅₀=14,8), or circular (IC₅₀=31,9) or parallel (IC₅₀=32,1), the latter having the highest IC₅₀. The proportional distribution showed that the parallel morphology related to a lower sensitivity to tridemorph was present in a higher percentage of individuals (37 %), followed by an undefined morphology (32 %) related to a higher yet intermediate sensitivity. The irregular morphology related to the highest sensitivity to tridemorph was found at a lower proportion (19 %), while a lower percentage of individuals (12 %) presented a circular morphology, characterized by having the second lowest levels of sensitivity, very close to those observed for the parallel morphology (Table 3y).

Population response. Among the factors there were no significant differences observed (Table 2y), although within the factors, when comparing averages, there were significant differences for the morphology of the mycelium. A population with a higher sensitivity to tridemorf was related to individuals with a circular (IC₅₀=1,54) or irregular morphology (IC₅₀=1,59), whereas a lower sensitivity to tridemorf was related to individuals with an undefined (IC₅₀=1,76) or parallel morphology (IC₅₀=1,87). In this case, proportional distribution showed that the parallel morphology related to a lower sensitivity was found in a greater proportion within the population (34%), followed by undefined (25%), irregular (22%), and circular morphologies (19%) (Table 3z).

Length of the Germ Tube and Biomass Production in Relation to Sensitivity to Fungicides

The length reached by the germ tube and the amount of biomass produced in vitro and in the absence of the fungicide did not display significant differences with the degree of sensitivity, acquired or inherent, of an individual or a population an individual belongs to, toward piraclostrobin, epoxiconazol, or tridemorf (Table 2z). In other words, a higher or lower IC₅₀ had no relation with a longer or shorter length of the germ tube or production of biomass (Table 1).

Length of the Germ Tube in Relation to Biomass Production

In the absence of the fungicide, the length of the germ tube and the amounts of biomass were compared, and no significant differences were found between these two factors (Table 2z), that is to say, a greater or less elongation of the germ tube had no relation with the production of more or less mycelia (Table 1).