MATERIALS AND METHODS

Sampling: The study was carried out during the spring-summer cycles of 2013 and 2014, in four commercial fields, two of which were located on the Coast of Hermosillo (CH1, CH2) and two in the Valley of Guaymas (VG1, VG2), in Sonora, Mexico, which account for 10% of the state’s surface used for the production of watermelon. The varieties planted in these fields were Sugar Red, SuperSeedless 7187HQ F1 and Precious Petit. In each cycle, 40 plants (10 plants per field) were collected at random in the areas with wilting and dryness of runners . The samples were placed in polyethylene bags, which were labelled and transported in containers with ice to the laboratory for processing.

Fungal isolation. The plant roots were washed using water, dried with paper towels, and cut into 1 cm pieces. Segments were taken from the crown, main root and secondary roots. They were disinfected by submerging them for 2 min in a solution prepared with sodium hypochlorite at 6%, ethyl alcohol at 96% and sterilized distilled water in a 1:1:8 ratio, respectivly. They were then placed in dishes with agar-water at 2%. The dishes were incubated at 25 ± 0.1 °C until the mycelium produced from the pieces of plant allowed for the extraction of a hypha tip. The hypha tips were planted successfully in Potato Dextrose Agar (PDA) supplemented with a solution of streptomycin/neomycin and incubated at 25 ± 0.1 °C, until a pure culture was obtained.

Morphological and cultural characterization. All isolates were grouped based on the characteristics of the cultures and the color developed on both sides of the PDA plate. A representative of each group was used for the morphological characterization. In the isolations with characteristics of Fusarium the presence and morphology of microconidia, macroconidia, and Chlamydospores was established from their growth in carnation leaf agar (CLA), after seven days in the dark at 25 ± 0.1 °C (^{Leslie and Summerell, 2006}). The isolations with characteristics of Rhizoctonia genus were identified by observing vegetative characteristics such as the color of the mycelium, septa, constrictions near the ramifications, during growth in PDA or Malt Extract Agar (MEA). To determine the number of nuclei, hyphae were stained with trypan blue in lactophenol. Growth rate was determined in PDA, keeping the cultures at 25 ± 0.1 °C and photoperiods of 14h/10h of light/darkness. The diameter of the culture was measured every 24 h until the mycelium covered the dish completely (^{Sneh et al., 1996}). All isolations were initially identified up to the genus level.

DNA Extraction. A mycelia from pure cultures in PDA were taken with a sterile microbiological spatula and placed in a tube of the Kit Power Soil DNA Isolation (MoBIO Laboratories, California, U.S.A.). Cell lysis was carried out in a Precellys Evolution Homogenizer (Bertin Technologies, France), stirring the tubes at 6500 rpm for three 20 s cycles with 20 s pauses. DNA integrity was verified in a 2% agarose gel. The extracted DNA was quantified in the NanoDrop 1000 (ThermoScientific). Samples with an absorbance ratio of 260/280 between 1.8 and 2 were amplified. The DNA was stored at -20°C until its use.

DNA amplification sequencing. The non-codifying region of the internal transcribed spacer (ITS) was amplified with a segment of the region that codified for the DNA-directed RNA polymerase II second largest subunit (RPB2) of all the isolations obtained. In addition, it was amplified the region that codifies for translation elongation factor 1-α (TEF1) for the isolates with Fusarium characteristics. Information about the primers used is shown in Table 1.

The polymerase chain reaction was carried out by mixing 12.5 ml of GoTaq^® Green Master Mix (Promega), 1 μl of each forward and reverse primer (IDT Technologies) 10 mM, 1 μl of DNA (1 ng/ μl)and molecular biology degree water up to 25 μl final volume. The PCR products were separated by electrophoresis in agarose gel at 2%, dyed with GelRed (Biotium Inc) in 5X Green GoTaq reaction buffer (Promega) 15 μl:1ml. They were visualized in UV light (DigiDoc-It™, UVP) observing the size of the amplicon and its purity. Purification was carried out using ExoSAP-IT PCR Product Cleanup (Affymetrix) or with by cutting bands of the expected size with the Wizard^® SV Gel kit and PCR Clean-Up System (Promega). Purified amplicons were sequenced in both directions with ABI 3730xl DNA Analyzer (Applied Biosystems) in GENEWIZ. Each sequence was reviewed manually and nucleotides in ambiguous positions were corrected with complementary sequences obtained with both primers, using the software ChromasPro v2.1.6. The sequences from regions ITS, RPB2 and TEF were compared by alignment with those contained in the databases Fusarium MLST (http://www.cbs.knaw.nl/Fusarium), Fusarium ID (http://isolate.fusariumdb.org) and in the National Center for Biotechnology Information (NCBI), in which the sequences from regions ITS and RPB2 of Rhizoctonia spp. were also compared, using the “Basic local alignment search tool” (BLAST) (^{Altschul et al., 1990}).

Phylogenetic analysis. The Fusarium species and the anastomosic groups of Rhizoctonia and Ceratobasidium were determined separately in two data matrices. In each case, several alignments were carried out using the software Clustal Omega (^{Sievers et al., 2011}). The sequences were linked and edited using the software UltraEdit32. The phylogenetic analysis of each data matrix was carried out separately under the criterion of Maximum Likelihood Estimation (MLE) with the software PAUP 4.0a152 (^{Swofford, 2002}). The best nucleotide substitution model was established using ModelTest (^{Posada and Crandall, 1998}).

The trees were viewed and modified in FigTree and exported to graphic editors. The consensus trees for Rhizoctonia spp. and Fusarium spp., were rooted using Botryobasidium simile (isolation GEL2348) and Neofusicoccum parvum (strain CCF216), respectively.

Pathogenicity tests. In compliance with Koch’s postulates, to assure that the isolates obtained were the causal agents of the root rotting in watermelon plants, a monosporic culture was chosen at random from each one of the species identified in the molecular analysis. The treatments were: 1) Non-inoculated control, 2) F. falciforme, 3) F. oxysporum, 4) F. brachygibbosum, 5) R. solani, 6) Ceratobasidium sp, 7) F. brachygibbosum + F. solani, 8) F. brachygibbosum + F. oxysporum, 9) R. solani + Ceratobasidium sp. 10) Ceratobasidium sp. + F. solani and 11) Ceratobasidium sp. + F. oxysporum. Healthy plants of the varieties SuperSeedless 7187HQ F1 and Precious Petit, aged 21 days, established in pots with a mixture of soil and perlite (1:3). Six plants were inoculated for each treatment. Each plant was treated with rotting and discoloration of vascular bundles, typical symptoms of damages by Fusarium spp. Four discs, 8 mm in diameter, were placed around the root. Healthy plants treated with sterile PDA discs were used as controls. The pots were placed in a controlled environment chamber at 25 ± 0.1 °C, with a photoperiod of 14h/10h day/night, until the appearance of symptoms. Irrigation was carried out based on water requirements, and a Hoagland nutrient solution was applied on a weekly basis. The number of disease plants was observed. The percentage of infected roots was determined using ten root segments from each plant. These tissue fragments were disinfected separately and placed in PDA. The evaluation was carried out observing the development of mycelia after seven-day incubation.

RESULTS

Damages observed during sampling. Figure 1A shows the damages observed in the crop. Plants with symptoms displayed different types of rotting in the crown, stem, roots and rootlets. One type of lesion was small, concave or not, brownish (Figure 1B) and moist-looking, typical in Rhizoctonia sp. In certain cases, this type of lesions was observed to have small pustules. In other plants, lesions were light brown, discolored and with vascular beam rot, typical in symptoms of damages by Fusarium spp. (Figure 1C).

Figure 1 Damage on planted watermelon crops. A) Damage on the field, B) Roots of diseased plants, C) Vascular damage. 

Isolation and morphology of the colonies and reproductive structures of fungi. A total of 45 fungal isolates were obtained from the four sampling sites, and in each one, at least one species of Rhizoctonia and Fusarium was isolated. The distribution of species by site is shown in Figure 2.

Figure 2 Distribution of pathogenic fungi species, by sampling sites. CH1= Coast of Hermosillo 1. CH2=Coast of Hermosillo 2. VG1=Valley of Guaymas 1. VG2=Valley of Guaymas 2. 

The morphological analysis helped to form two groups of isolates. One group was formed of 13 isolations that presented colonies with abundant aerial hyphae, ivory-colored at first, turning light brown or brown after 7 days. In these isolations, hyphae were robust, with branches in right angles, constriction of the ramification and the formation of a septum near the point of origin, without spores, all of which are typical characteristics of Rhizoctonia sp. (^{Sneh et al., 1996}). Average growth rate was 1.25 mm h^-1. Nuclei staining revealed the presence of isolates with both binuclear and multinuclear cells (Figure 3).

Figure 3 Morphology of the colonies and colors of the medium in PDA of Rhizoctonia spp., Ceratobasidium sp., F. falciforme, F. oxysporum, F. brachygibbosum, (1A-5A and 1B-5B) respectively. (1C) Polynuclear hyphae. (1D) Septa. (2C) Binuclear hyphae. (2D) Monilioid cells. (3C) Monophialides and microconidia. (3D) Macroconidia. (4C) Microconidi in situ in CLA. (4D) Microconidia. (5C) Macroconidia. (5D) Clamidospores. 

The second group of 32 isolates showed the formation of macroconidia, microconidia, chlamydospores and monophyalides in CLA, typical of Fusarium sp., according to descriptions by ^{Leslie and Summerell (2006)}. Twenty-five isolations fit the description of F. solani: cream-colored colonies with red to dark gray pigments in the obverse oval-shaped microconidia without septa, long monophyalides, single or paired chlamydospores, abundant straight macroconidia with 3 to 5 septa. Other 2 presented typical characteristics for F. oxysporum: cottonlike mycelia, scarce white to pale violet and purple color in the agar; short monophyalides. The rest of the isolates produced white mycelia, which turned pink, with yellow sporodochia; oval unicellular microconidia, produced in monophyalides, curved macroconidia with 3 to 5 septa, with wide central cells, slightly pointy apex, single or chained chlamydospores (Figure 3). No isolations were found with different morphology to Rhizoctonia or Fusarium.

Molecular identification and phylogenetic analysis of fungi. A first BLAST analysis from the region between the internal transcribed spacers I and II (ITS1-ITS2) of all the isolates helped determine that 25 sequences had a similarity of 99-100% with F. solani; 5 to F. brachygibbosum, 2 to F. oxysporum, 10 to Ceratobasidium sp. and 3 sequences to Rhizoctonia solani also with a similarity of 99-100% to NCBI homologous sequences.

The phylogenetic analysis under the ML criterion for the Fusarium genus was carried out with the linked matrix of genes ITS, RPB2 and TEF1 of the 32 isolations of this study and of 25 reference strains deposited in culture collections (^{Al-Hatmi et al., 2016}). The best nucleotide substitution model was TIM2+I+G. Total of 1816 nucleotides were considered in the data set. The multilocus analysis helped define the correct identity of the isolations initially proposed as F. solani, since they form a separate clade with a similarity of 100 % with the type isolations of F. falciforme, a member of the Fusarium solani species complex (FSSC). The identity of F. brachygibbosum and F. oxysporum was corroborated (Figure 4).

Figure 4 Phylogeny by MV of the chained sequences of the genes ITS, EF and RPB2 of 8 Fusarium spp. isolations (in bold) representative of 29 obtained from watermelon plants with wilting and root rot. Bootstrap of 1000 replications; external group: Neofusicoccum parvum. 

A similar analysis was carried out with the sequences of isolations from Rhizoctonia and Ceratobasidium sp. a data matrix was integrated by the regions ITS and RPB2 of 49 isolations, including those from the present study and 36 references strains (Cantharellales), obtained from culture collections (^{González et al., 2016}). The best nucleotide substitution model was GTR+I+G. Total of 1355 nucleotides were considered from each isolation in the data set. This analysis established the identity of two anastomosic groups for the genus Rhizoctonia, AG-4 and AG-6 and two for the genus Ceratobasidium, AG-A and AG-F (Figure 5). Table 2 shows the identity and the accession numbers of the sequences obtained.

Figure 5 Phylogeny by MV of the chained sequences of the genes ITS, EF and RPB2 of 8 Rhizoctonia spp. isolations (in bold) representative of 16 obtained from watermelon plants with wilting and root rot. Bootstrap of 1000 replications; external group: Botryobasidium simile. AG=Anastomotic group.  

Pathogenicity tests: The appearance of symptoms in plants took place 14 days after inoculation, in the form of lesions in the roots and the base of the stem. All the isolation and their combinations caused the death of the plants after 21 days. Table 3 shows the percentage of infected roots in each treatment. Control plants presented no symptoms.

DISCUSSION

Although Sonora is the main watermelon producing state in Mexico, there are no formal reports on the fungi related to rotting. The pathological data in this investigation show that, individually or in a group, at least five different fungal species caused root rot, which led to death of watermelon plants before they reached physiological maturity, during the formation and development of fruits. The characteristics of the colonies, morphology, nuclei staining and phylogenetic analysis of sequences helped to identify two species of the Ceratobasidiaceae family: Rhizoctonia solani and Ceratobasidium sp., and three of the Fusarium genus: F. falciforme, F. oxysporum and F. brachygibbosum, showing that there is a diverse community of fungi causing root rot in watermelon planted in the Coast of Hermosillo and the Valley of Guaymas, Sonora.

The predominant species was found to be F. falciforme, with 25 isolations from three fields of the two sites sampled. Initially, this fungus was identified as the polytypic morphospecies F. solani. Based on the linked phylogenetic analysis of the regions ITS, TEF-1α and RPB2 using ML, a separate clade is formed within the Fusarium solani Species Complex (FSSC). This complex groups at least 60 different species that, because of the similarity in the morphology of their conidia, are known as cryptic species. They have a wide range of hosts and have been subdivided into formae speciale, depending on the specificity of the host (^{O’Donnell et al., 2015}). Recent phylogenetic analyses have revealed that each formae speciale belongs to a biologically and phylogenetically different species (^{Coleman, 2016}; ^{O’Donnell et al., 1998}). F. falciforme has been reported as a causal agent of wilting and root rot in lima bean plants and chickpea in Brazil (^{Sousa et al., 2017}, ^{Cabral et al., 2016}).

Five F. brachygibbosum isolations belonging to Fusarium sambucinum Species Complex (FSSC) were identified in three sampling sites. The characteristics of the colonies coincide with the first description published (^{Padwick, 1945}), they present abundant white to pink aerial mycelia white to amber sclerotia of up to 2.0 mm in diameter, oval or fusiform conidia, hyperbolically curved macroconidia, terminal or alternated chlamydospores, single or in chain, generally unicellular. As a preliminary product of this investigation, F. brachygibbosum was reported for the first time as a pathogenic agent, causing this wilting in watermelon seeds (^{Rentería-Martínez et al., 2015}). It has recently been recorded as a causal agent of rotting in the stem of maize plants (^{Shan et al., 2017}), wilting and regressive death in Euphorbia larica and olive trees (^{Al-Mahmooli et al., 2013}; ^{Trabelsi et al., 2017}) and of rotting and cankers in almond and chestnut trees (^{Stack et al., 2017}; ^{Marek et al., 2013}).

F. oxysporum f. sp. niveum belongs to the Fusarium oxysporum Species Complex (FOSC); it was only isolated from the fields located in the Valley of Guaymas. The components of FOSC cause vascular wilting and root rot in over 100 different plant species, and based on their specificity with the host, there have been reports of over 80 formae speciale. In this sense, an accurate diagnosis, even before symptoms appear, is crucial for the management of crops (^{López-Mondéjar et al., 2012}).

Regarding the Ceratobasidiaceae family, binuclear Rhizoctonia (teleomorph: Ceratobasidium sp.) was the species with the widest distribution, since it was isolated from all the fields sampled, and most frequently, in field No. 2 of the Coast of Hermosillo. The symptoms observed in roots and stems of infected plants were reddish-brown lesions located and slightly sunken in the base of the stem and 0.2 to 2.0 cm in length. In some cases, the discoloration presented by diseased plants affected almost 90 % of the root system (^{Meza-Möller et al., 2014}).

The phylogenetic analysis by ML showed that the isolations obtained from Ceratobasidium sp. belong to two different anastomosic groups: AG-A and AG-F, previously found in diseased Ipomoea batatas and Arachis hypogaea plants, respectively. The phylogenetic relations of Rhizoctonia fungi show consistency between the clades formed and the anastomosic groups to which they belong (^{González et al., 2016}).

It is important to point out that most of the Ceratobasidium sp. isolates found in this investigation belong to group AG-F, which coincides with earlier works on root rot in watermelon carried out in Arizona, U.S.A (^{Nischwitz et al., 2013}) and Italy (^{Aiello et al., 2012}). Ceratobasidium AG-F is also the causal agent of root rot in strawberry (^{Sharon et al., 2007}), Tagetes erecta (^{Saroj et al., 2013}) and pistachio (^{Alaei et al., 2017}).

There is no previous information on the pathogenicity of the anastomosic group AG-A in watermelon, but there is on sugar beet and apple trees (^{Wang and Wu, 2012}).

The least frequent isolations belong to the genus Rhizoctonia, which, due to the high variability in its geographic distribution, morphology, specificity of hosts, and pathogenicity, has also been proposed as a species complex (^{González et al., 2006}). Isolation RhDAG18 obtained from site CH2 belongs to AG-4 and has been morphologically delimited as Thanatephorus praticola (^{Mordue et al., 1989}), which was later corroborated by the multilocus phylogenetic analysis (^{González et al., 2016}). T. praticola has been presented in association with Ceratobasidium sp. AG-F as part of a fungal complex that cause root rot and deterioration in watermelon plantations in the production stage in Italy (^{Aiello et al., 2012}).

The pathogenicity tests showed that isolated fungi are causal agents of root rot in watermelon plants. The results of the percentage of infected roots showed that both watermelon varieties evaluated are susceptible to the five pathogens identified, inoculated separately or in combinations. In this regard, ^{Aiello et al. (2012)} detected percentages of incidence of the disease higher than 81 % when evaluating 6 different isolations of Ceratobasidium from the anastomosic group AG-F.

Due to economic losses resulting from root diseases in the areas of study, direct planting in the soil has been replaced almost entirely with watermelon grafted on patterns resistant to root rot. The commonly used rootstocks in the region are hybrids between Cucurbita maxima x C. moschata, although these hybrids have shown susceptibility to root rot (^{López-Elías et al., 2010}). Some rootstocks have shown to be susceptible to Fusarium in Spain, the main producer of cucurbits in Europe (^{Armengol et al., 2000}). On the other hand, the apparent resistance to wilting by Fusarium of some commercial varieties of watermelon depends on the type and initial concentration of the inoculant (^{Martyn and McLaughlin, 1983}); in this sense, the timely identification of the pathogen will help make a better choice of rootstock. Additionally, once the pathogenic species present has been accurately established, the quantification of its concentration will manage the disease better. Using the quantitative PCR (qPCR) technique, it is possible to detect up to one 1 pg of fungal DNA in plants without symptoms (^{Haegi et al., 2013}).

CONCLUSIONS

In the present study, Fusarium falciforme and Thanatephorus praticola are reported for the first time as causal agents of watermelon plants root rotting in Mexico. The Multilocus analysis corroborated Fusarium brachygibbosum and Ceratobasidium sp. identity, previously reported in the literature. This research is a first attempt at generating data that could lead to broader studies that may contribute to establish adequate control strategies for disease management in commercial watermelon plantations.