Materials and methods

The sites and methodology used to sample plants showing symptoms, isolation, monosporic cultures, DNA extraction, as well as to do the molecular identification and conduct pathogenicity tests on the isolates following Koch’s postulates, were described and published by ^{Rentería-Martínez et al. (2018}). A total of 25 Fusarium falciforme isolates, five F. brachygibbosum isolates, and two F. oxysporum isolates were identified.

Morphotype classification. Morphotypes of each species were established based on the type of growth, mycelium color and culture medium pigmentation. Assessments were conducted on one isolate that was representative of each morphotype.

Morphology and growth rate of the colonies. The macroscopic characteristics of the colonies (mycelium growth rate, color of the colony and pigmentation) were evaluated in potato-dextrose-agar (PDA) medium. The growth rate of each isolate was determined by taking a disk 8 mm in diameter containing 7-day-old colonies in PDA and placing it in the middle of another dish containing PDA. The dishes were incubated at 27 °C and the colony’s diameter was measured in two perpendicular directions every day until the dish was completely covered by the culture. Five replicates of each isolate were evaluated. The morphology of the colonies and the aerial pigmentation of the mycelium were recorded (^{Nelson et al., 1983}).

Analysis of microscopic characteristics. The different morphotypes were cultured in carnation leaf agar (CLA), according to the method of ^{Fisher et al. (1982}). To promote the formation of chlamydospores, soil-agar (SA) was used, according to the method of ^{Klotz et al. (1988}). The CLA and SA cultures were incubated under the same conditions: 25 °C, 12 h light/12 h darkness for two weeks.

Microscopic preparations on slides were observed using a compound light microscope (VELAB^TM). To observe and measure the size of the conidia, phialides and chlamydospores, pictures were taken at 40X and 100X, which were then analyzed using TSView software, version 6.2.4.5. The length and width of 50 macroconidia and 50 microconidia randomly selected from each isolate were measured.

Statistical analysis. Data on macroconidial and microconidial length and width were subjected to an analysis of variance (ANOVA) and media separation using Tukey’s test (α<0.05) and the JMP program v5.0.1a.

Results

Morphotype classification and growth rate. The 25 F. falciforme isolates were grouped in three morphotypes (Ff-14, Ff-49 and Ff-50), the five F. brachygibbosum isolates in four morphotypes (Fb-8, Fb-9, Fb-41 and Fb-67), and the two F. oxysporum isolates in one morphotype (Fo-38). Table 1 shows the growth rates and macroscopic characteristics observed in F. brachygibbosum, F. falciforme and F. oxysporum morphotypes grown in PDA. Figure 1 shows the presence of the four F. brachygibbosum pathogenic morphotypes found in watermelon fields in Sonora.

Microscopic characteristics. Table 2 summarizes the chlamydospores morphology and conidiogenic cells observed in isolates representative of each morphotype in the three species. In all cases, spherical, intercalary and terminal chlamydospores were observed, mainly singly and in pairs; however, the walls of the chlamydospores of F. brachygibbosum morphotypes (Figure 2A) are thicker, which seems to be a characteristic that distinguishes them from those of F. falciforme and F. oxysporum (Figure 2B-C). Regarding conidiogenic cells, F. falciforme morphotypes (Figure 2B) differ from the others because their monophialides are longer and thinner.

Table 3 shows a morphological description of the macroconidia and microconidia of isolates that are representative of each morphotype. There were some differences among the three species. The four F. brachygibbosum morphotypes had thin, straight macroconidia with three to five septa, the cells in the middle were slightly wide and the intermediate cells were curved towards the basal part. The apical cells were slightly hook-shaped, while the basal cells had a typical shape or were slightly foot-shaped (Figure 3). F. falciforme macroconidia were wide and straight, or slightly curved, but the ventral and dorsal planes were parallel along most of their length, the apical cells were rounded, while the basal cells were straight and cylindrical with a barely trimmed or round end (Figure 4B). The shape of F. oxysporum macroconidia was similar to that of F. brachygibbosum macroconidia, but their basal cells were pointed and lower in number, with a barely visible foot shape (Figure 4C).

F. brachygibbosum microconidia were oval, sometimes ovoid, usually with one septum, rarely with two or none. F. oxysporum microconidia were elliptical, oval and kidney-shaped with no septa. F. falciforme had oval, ellipsoid and kidney-shaped microconidia with a truncated basal part, and some of them were pyriform and fusiform with one or two septa, or none (Table 3, Figure 4 D-F). Due to shape variability, it was not possible to observe differences among macroconidia from different species, and only the absence of septa in all F. oxysporum (Figure 4F) microconidia observed seems to be a distinctive trait.

Figure 1  Morphological characteristics of colonies of four Fusarium brachygibbosum morphotypes grown in PDA. A = Front B = bottom.  

Figure 2 A-C: Morphology of chlamydospores of A = F. brachygibbosum; B = F. falciforme; and C = F. oxysporum. D-F: Morphology of conidiogenous cells: D = polyphialides and monophialides of F. brachygibbosum, E = monophialides of F. falciforme, and F = monophialides of F. oxysporum.  

Figure 3  Morphology of macroconidia of four F. brachygibbosum morphotypes. A = slightly hook-shaped apical cells; and B = slightly foot-shaped basal cells (Fb-8 and Fb-67) and typical foot-shaped cells (Fb-9 and FB-41).  

Table 4 shows the average length and width of the macroconidia, according to the number of septa and the average of 50 cells. The statistical analysis detected significant differences in the average length and width of the macroconidia among the isolates, but due to the variability in size within the species, it was not possible to observe the differences between them. Macroconidia with Fb-8 and Fb-9 morphotypes had no significant size differences compared to F. oxysporum Fo-38 morphotypes. No significant differences were detected in the size of Fb-41, Fo-39 and Ff-49 macroconidia, nor among Fb-67, Ff-50 and Ff-14 macroconidia.

Figure 4 Morphology of macroconidia and microconidia of three Fusarium species pathogenic to watermelon. A = mac roconidia of F. brachygibbosum, B = macroconidia of F. falciforme, and C) macroconidia of F. oxysporum. D = macroconidia of F. brachygibbosum, E = microconidia of F. falciforme, and F) microconidia of F. oxysporum.  

Table 5 shows the average length and width of the macroconidia, according to the number of septa and the average of 50 cells. In general, significant differences were detected in the size of the macroconidia among isolates but not among species. The Fb-41, Fb-67 and Fb-9 morphotypes had the longest microconidia, while Fb-8 had the shortest.

Discussion

^{Padwick (1945}) stated that F. brachygibbosum grows in PDA and develops an abundance of aerial white mycelia, pinkish tone, and blood-red color in the agar; its microconidia are typically ovoid to fusiform, or slightly curved when they have septa, with one, two, three or no septa. Hyperbolically curved macroconidia are finely dispersed over the agar surface; they have wide central cells and 0-5 septa. Sharp apical cells and slightly sharp basal cells have a typical foot-like shape; terminal and intercalary chlamydospores, single or in chains, are usually unicellular and globose, occasionally with two smooth and granular cells. This description is in agreement with the four F. brachygibbosum morphotypes that were found in this study, but the color of the colony and the medium coincides only with the color of Fb-67 and not with the color of Fb-8, Fb-9 and Fb-41.

In general, the type of mycelium, the color of the colonies and the culture medium, as well as the morphology of the macro- and microconidia observed in the most recent F. brachygibbosum studies (^{Al-Sadi et al., 2012}; ^{Mirhosseini et al., 2014}; ^{Renteria-Martinez et al., 2015}; ^{Stack et al., 2017}; ^{Shan et al., 2017}; ^{Xia et al., 2018}) correspond to the morphotype described by ^{Padwick (1945}). All these studies found white colonies that turned yellow with pink or orange tones, and stained the PDA medium bright red. When they were grown in CLA, the macroconidia had thick walls, a bulky center and slightly curved ends (attenuated hook), and a typical foot-like or slightly foot-shaped basal cell; most of them had 4-5 septa.

The F. oxysporum morphotype developed white aerial mycelium that turned purple after a week in PDA, a result similar to that described by ^{Kleczewski and Egel (2011}) regarding F. oxysporum f.sp. niveum race 1, but with no sporodochia formation. When cultured in CLA, the characteristics of the chlamydospores, conidiogenic cells, macroconidia and microconidia were in agreement with the results obtained by ^{Leslie and Summerell (2006}) for the species.

The type of growth colonies in PDA makes it possible to distinguish between F. brachygibbosum (aerial) and F. falciforme (prostrate) morphotypes. The presence of long and thin monophialides observed in F. falciforme morphotypes (^{Summerbell and Schroers, 2002}) is a characteristic that allows differentiating them from F. brachygibbosum and F. oxysporum, which develop shorter monophialides. The shape of F. falciforme macroconidia is another trait that distinguishes it from the other species. Similar to the results obtained by ^{Chehri et al. (2015}), macroconidia of this species were polyseptate, wide and straight or slightly curved, with round apical cells and basal cells with slightly trimmed ends.

The color developed in the agar, the type of conidiogenic cells, as well as macroconidial morphology and size, can cause F. brachygibbosum to be mistakenly identified as F. oxysporum, which is more common in most of the crops in Sonora, but the morphology of the chlamydospores, specifically the thickness of their walls, may be a distinctive characteristic that allows differentiating the two species. The morphology of F. brachygibbosum chlamydospores is in agreement with the results obtained by ^{Cao et al. (2018}). Another characteristic that differs among these species is the lack of septa in F. oxysporum microconidia, while F. brachygibbosum had mostly microconidia with one septum.

The variability in macroconidia and microconidia size among morphotypes of the same species did not allow differentiating among the Fusarium species.

F. brachygibbosum conidia of different sizes have been reported (^{Xia et al., 2018}; ^{Shan et al., 2017}; ^{Mirhosseini et al., 2014}; ^{Al-Mahmooli et al., 2013}) and, in most cases, they have been shorter than the ones observed in this study. The size of F. falciforme macroconidia was similar to the size reported by ^{Chehri et al. (2015}). Also, the morphology and size of F. oxysporum macroconidia were in agreement with the results obtained by ^{Chehri et al. (2011)}.

Conclusions

Some characteristics, such as the thickness of chlamydospore walls, macroconidia morphology, type of conidiogenic cells and the presence or absence of septa in microconidia can be distinctive and useful for identifying F. brachygibbosum, F. falciforme and F. oxysporum that are pathogenic to watermelon plants grown in Sonora, Mexico. However, the identified morphologies need to be corroborated through molecular analysis.