Materials and methods

Collection of samples and obtaining monosporic isolates. Symptomatic soybean plant leaflets were collected in the del Carrizo and del Fuerte valleys between September and October, 2018 (Cuadro 1). Samples were taken from eight commercial soybean var. Nainari fields. These varied between 10 and 30 ha of crops, five in the stage of grain filling (R6) and three in the stage of the beginning of maturity (R7) (^{Fehr and Caviness, 1977}); in each lot, the sampling method used (five of gold) was to select five plants, out of which five leaflets, each with light to dark brown foliar spots surrounded by a yellow halo were taken. The isolates of fungi related to symptomatic leaflets were carried out in agar water (AA; Bioxon; Cuautitlán Izcalli, State of Mexico, Mexico) following procedures described by ^{Maraite et al. (1997}). They were incubated at 25 ± 2 °C for 72 h, until mycelial growth was observed. To purify the isolates, hypha tips were transferred to potato-dextrose-agar (PDA; Bioxon; Cuautitlán Izcalli, State of Mexico, Mexico).

The monosporic isolates were obtained by serial dilutions, which were planted in the AA at 1.6% and incubated for 4 h; a germinated conidium was transferred from there to a PDA medium and the plants were incubated at 25 °C. Monosoporic isolates were stored at 15 °C in tilted test tubes with PDA, with added mineral oil (^{Flores-Córdova et al., 2013}).

Phenotypic characteristics of mycelial cultures. The phenotypic characteristics of the cultures were determined with PDA. The cultures were incubated for 11 days at 25 °C, with a 12 h light-darkness regime, and the mycelial growth rate was determined (^{Kim et al., 2005}), along with the color of the culture, using standard ^{Kelly and Judd colors (1976}).

Morphometry of the conidia of isolates. Fragments, 5 mm in diameter, were cut from the margin of the four-day-old cultures in PDA and transferred to the same medium. The shape and color of the septa, as well as the length and width of 40 conidia from each isolate were determined seven days after planting in the same medium and incubated at 25 °C. Preparations were then made with lactophenol blue to be measured in a micrometer installed in a compound microscope (Labmed; Lab America, Inc. Los Angeles, CA, USA).

Extraction of DNA, PCR and sequencing. The extraction of genonic DNA was carried out using 50-100 mg of mycelia in seven-day-old PDA cultures, using DNAzol as a reagent (Molecular Research Center Inc., Car. No. DN127, Cincinnati, OH, USA), following the instructions by the manufacturer. The quality and concentration of the DNA extracted was determined in a NanoDrop 2000c spectrophotometer (Thermo Scientific, Wilmington, DE). To confirm the identity of the isolates, the gene from the glyceraldehyde-3-phosphate dehydrogenase (gapdh) was amplified using primers gpd1 (5’-CAA CGG CTT CGG TCG CAT TG-3’) and gpd2 (5’-GCC AAG CAG TTG GTT GTG C-3’), which amplify a fragment of approximately 580 pairs of bases (^{Berbee et al., 1999}). The PCR’s reaction took place in a final volume of 50 mL, which contained 1 mL of genomic DNA, GoTaq Flexi 1X Buffer, 1.25 mM of MgCl2, 0.2 mM of each primer, 0.2 mM of dNTP´s and 1 U of DNA polymerase GoTaq Flexi (Promega, Madison, WI, USA). Amplification conditions were as follows: initial denaturalization at 95 °C for 2 min, 35 denaturalization cycles at 95 °C for 1 min, alignment at 60 °C for 1 min, an extension at 72 °C for 45 s and a final extension at 72 °C for 10 min. The amplified fragments were visualized in a 2% agarose gel stained with ethidium bromide in a Chemidoc (BIO-RAD) and the bands of interest were split for their later purification using the QIAquick Gel Extraction kit (Qiagen, Hilden, Germany). The quantification od the purified PCR products were carried out with a NanoDrop 2000c spectrophotometer. The fragments were sequenced unidirectionally with the primer gpd1 in an AB3730 (Applied Biosystems, Foster City, CA, USA), in the genomic services lab of the National Genonic Laboratory for Biodiversity (LANGEBIO).

Phylogenetic analysis. The sequences were edited in the software BioEdit version 7.0.5.3 (^{Hall, 1999}) and compared with other sequences deposited in the GenBank data base, using the BlastN algorithm. The sequences were aligned along with a set of 68 references sequences of the genus Curvularia and a sequence of Bipolaris maydis as an external group (^{Madrid et al., 2014}; ^{Manamgoda et al., 2015}; ^{Tomaso-Peterson et al., 2016}; ^{Mehrabi-Koushki e}t al., 2018; ^{Tan et al., 2018}; ^{Kiss et al., 2019}), using the MUSCLE aligner (^{Edgar, 2004}) implemented in MegaX (^{Kumar et al., 2018}). The phylogenetic analysis was carried out in MegaX, using the maximum likelihood method and the Hasegawa-Kishino-Yano (HKY) model, with a gamma distribution (four categories) and unvariable sites. All the gaps were considered in the analysis and the robustness of the topology was evaluated with 1000 bootstrap replications. Finally, the phylogram was edited using the software FigTree version 1.4.0 (^{Rambaut, 2014}).

Pathogenicity tests in a greenhouse. To determine the pathogenicity of the eight isolates associated with the disease, soybean plants var. Nainari were used, which were planted in 5L plastic pots with a pasteurized substrate (sand 48.0, silt 19.4 and clay 32.5%, respectively and a pH of 7.0); one seed was placed in each pot. The plants were watered according to each one’s requirements and fertilized every week with Miracle-Gro (The Scotts Company LLC; Marysville, Ohio, USA), following the manufacturer’s recommendations. Four replicates were used for each isolate (four pots with one plant each).

Preparation of inoculum. The monoconidial isolates were developed in PDA at 25 °C, with a 12 h light-darkness regime for 10 days. The conidia were obtained by adding 10 mL of sterile distilled water to each culture and scraping with a sterilized spatula to release them. The conidial suspensions were filtered through a double cheese mesh to remove the mycelia (^{Prieto et al., 2016}) and the inoculum concentrations of each isolate were adjusted from 5 x 10⁴ to 7 x 10⁴ conidia mL^-1 with a hematocytometer.

The conidial suspensions were sprayed on plants (in the R6 stage) to the point of run-off. The control plants were sprayed with sterile distilled water. The plants were then covered for 24 h with 90 x 20 cm plastic bags with paper towels saturated with distilled water to achieve a relative humidity of 85-90% in their interior and they were kept at 22 to 28 °C; they underwent the same moisture and temperature conditions for 12 h a day, every day for seven consecutive days. The plants then remained for 21 days on the greenhouse benches, where temperatures ranged from 20 to 24 °C with relative humidity ranges of 50 to 75%. The inoculated and control plants were distributed in a totally random design. The pathogenicity of the isolates was determined 30 days after inoculation. Ten leaflets were collected at random from each plant and the percentage of the leaf area affected (LAA) was determined, following procedures described ^{Soares et al. (2009}). To comply with Koch’s postulates, after the experiments, isolates were carried out from symptomatic leaflets taken from inoculated plants. The identity of the isolates was confirmed based on the morphology of the cultures and the morphometry of twenty conidia from each isolate.

Statistical analysis. The percentages of LAA were transformed using the arcsine function (^{Gómez y Gómez, 1984}) before performing an ANOVA. The comparison of means was carried out with Tukey’s test (^{Little and Hills, 1973}), using the statistical package SAS^® version 9.4 (1999). The experiments were held twice. Given that the statistical analyses indicated a significant interaction between isolates and the experiments, the results of both experiments are presented.

Results and discussion

Cultural and morphometric characteristics of fungal isolates. The phenotypic characteristics of the cultures, the conidiophores and conidia of eight isolates obtained during the sampling period (Table 1) were determined in PDA after eight days of incubation. This helped locate the isolates in the genus Curvularia (^{Manamgoda et al., 2015}) genus. No fungal teleomorph was found during the sampling period, although the genus Cochliobolus has been confirmed as the teleomorph of some species of Curvularia (^{Marin-Felix et al., 2020}). The phenology of the colonies, the characteristics of conidiophores and the morphometry of the Curvularia species are described below.

Isolates Cv-3, Cv-5 and Cv-8. Cultures with a mycelial growth rate of 3.7 - 4.4 mm/day on PDA, with a velvet-like aspect, olive-gray, with faint concentric circles (Figure1A); reverse side of the cultures with olive-brown pigmentation (Figure 1B). Straight, dark brown septate conidiophores, geniculated in the terminal part (Figure 1C); conidia, 15.0 - 27.5 μm in length by 7.5 - 15.0 μm in width, with 2 to 4 transversal septa, the middle septum being the bulkiest one and with the most pigmentation, a pronounced curvature and a dark hilum (Table 2; Figure 1D). The morphometric characteristics of these are similar to those recorded for C. muehlenbeckiae (^{Berbee et al., 1999}; ^{Madrid et al., 2014}).

Isolate Cv-4. Cultures with a cotton- or suede-like appearance, with a mycelial growth rate of 4.0 mm/day on PDA, colored dark gray with uneven tones of brown to light gray (Figure 1E) and the front, dark golden brown in color (Figure 1F). Light-brown conidiophore and geniculated on the top part. Cylindrical conidia, with three to four distosepts, ellipsoidal, golden maroon in color (Figure 1G and H), 10.0 - 22.5 µm in length by 7.5 - 10.0 µm in width (Table 2). The cultural and morphometric characteristics are similar to those found in C. spicifera (^{Ayoubi et al., 2017}; ^{Qostal et al., 2019}).

Isolate Cv-11. Olive-gray cultures, scarcely dense, with a velvetlike aspect, very noticeable concentric rings (Figure 1I) and a mycelial growth rate of 3.9 mm/day on PDA. The reverse side of the cultures dark to olive green (Figure 1J). Isolated and geniculated conidiophore in the terminal part, without ramifications (Figure 1K). Conidia between 16.2 and 27.5 µm in length by 6.2 to 10.0 µm in width (Table 2), with a smooth wall, ellipsoidal to cylindrical, rounded tips, with 2 to 4 transversal distosepta (Figure 1L), protuberant hilum (Arrow in Figure 1L), coinciding with the characteristics described for C. ahvazensis (^{Mehrabi-Koushki et al., 2018}).

Isolates Cv-19, Cv-21 and Cv-29. Cultures with a mycelial growth rate of 2.8 - 4.3 mm/day in PDA, with a suede-like aspect, undefined concentric rings and whitish plumes (Figure 1M), and a dark maroon to black reverse side (Figure 1N). Conidiophore are erect to flexible, septated, often geniculated (Figure 1O). Conidia are 10 to 35 µm in length by 5 to 12.5 µm in width, ellipsoidal or rounded on the tips, pale to dark maroon, 2 a 5 distosepts (generally 4), smooth conidial wall (Table 2; Figure 1P y Q). The phenotypical characteristics of the cultures, as well as the morphometry of the conidia are similar to those in species phylogenetically related to C. hawaiensis (^{Manamgoda et al., 2015}).

Figure 1 Morphology species of Curvularia in PDA related to leaf spot in soybean. A-D) Representative morphology of isolates Cv-3, Cv-5 and Cv-8 (C. muehlenbeckiae); A) Front side of the culture; B) Reverse side of the culture; C-D) Conidiophores and conidia. E-H) Isolate Cv-4 (C. spicifera); E) Front side of the culture; F) Reverse side of the culture; G and H) Conidiophores and conidia. I-L) Isolate Cv-11 (C. ahvazensis); I) Front side of the culture; J) Reverse side of the culture; K-L) Conidiophores and conidia; L) Conidia. M-Q) Representative morphology of isolates Cv-19, Cv-21, Cv-29 (C. hawaiensis); M) Front side of the culture; N) Reverse side of the culture; O-P-Q) Conidiophore and conidia. Scale bar=20 µm. 

Molecular identification. The partial sequences of the gapdh gene were deposited in the GenBank database (accession MT680146-MT680150 and MT680152-MT680154). The analyses confirmed the identity of the isolates at a genus level. In turn, the comparison of the sequences in the GenBank database displayed high percentages of similarity with diverse Curvularia species; isolates Cv-3, Cv-5 and Cv-8 were similar to C. muehlenbeckiae strain UTHSC 08-2905 (accession LT715807; 99% identity) and CBS 144.63 (accession LT715806, 100% identity); isolate Cv-4 was similar to C. spicifera strain CBS 198.31 (accession LT715814.1, 99% identity); isolate Cv-11 recorded high percentages of similarity with C. hawaiiensis strain CBS 173.57 (accession LT715812, 99% identity), C. nodosa strain CPC 28812 (accession MF490840, 98.5% identity), C. perotidis strain CBS 350.90 (accession LT715813, 97.9% identity) and C. dactyloctenii strain BRIP 12846 (accession KJ415401, 98.5% identity); whereas isolates Cv-19, Cv-21 and Cv-29 shared high percentages of similarity with C. beasleyi strain BRIP 15854 (accession MH433639, 99.1% identity), C. dactyloctenii strain BRIP 12846 (accession KJ415401, 98.9% identity), C. hawaiiensis strain CBS 173.57 (accession LT715812, 98.7% identity), C. nodosa strain CPC 28812 (accession MF490840, 98.5% identity) and C. perotidis strain CBS 350.90 (accession LT715813, 97.9% identity).

Figure 2 shows the phylogram inferred from the partial sequence of gene gapdh. Isolates Cv-3, Cv-5 and Cv-8 were grouped, along with the C. muehlenbeckiae type sequence, with a high bootstrap support (90.9%), which coincides with the morphological description of these isolates for the species at hand (Figure 1A-D). Isolate Cv-4 was identified as C. spicifera, since it grouped with the type sequence of this species (bootstrap 94.1%) and its morphological characteristics coincide with those reported previously. Isolate Cv-11 is grouped with C. ahvazensis (52.1% bootstrap) and its morphological characteristics coincide with those reported for the species. The phylogram shows that isolates Cv-19, Cv-21 and Cv-29 are different and do not correspond to any of the species of Curvularia used as references in the analysis, and were therefore considered as Curvularia sp. The three isolates are phylogenetically near to C. beasleyi, the species with which they share 99 - 99.2% similarities.

This group of fungi has a high morphological complexity and diversity of cryptic species; therefore, its taxonomy has not been entirely solved (^{Manamgoda et al., 2015}). Currently, the genus covers just over 100 described species, out of which only ~80 have been accepted and their taxonomic locations have been established through multigene phylogenetic analyses (^{Manamgoda et al., 2015}; ^{Tan et al., 2018}; ^{Kiss et al., 2019}). However, the gene gapdh on its own can solve the Curvularia species (^{Manamgoda et al., 2015}), and it is the locus with the most phylogenetic information among the loci analyzed for this fungal genus (^{Ferdinandez et al., 2019}). For isolates Cv-19, Cv-21 and Cv-29, the identities of which could not be determined, we recommend carrying out a more detailed molecular characterization to corroborate them as a new lineage within the genus and give them a name. In order to do this, the sequencing of the ITS region and/or the translation elongation factor (TEF) gene must be included, as it has been done previously in the registration of new species of Curvularia (^{Heidari et al., 2018}; ^{Mehrabi-Koushki et al., 2018}).

Pathogenicity test. All the isolates obtained from symptomatic soybean leaflets were pathogenic in soybean var. Nainari and caused similar symptoms to those observed in the field (Figure 3). The first symptoms were found in the basal third and middle of the plants six days after inoculation, while defoliation began 15 days after inoculation; in turn, the symptoms caused by the different Curvularia species were similar, but with variations in the degree of virulence. In experiment 1, 30 days after inoculation, the plants inoculated with the C. ahvazensis (Cv-11) and Curvularia sp. (Cv-21) isolates presented an LAA of 30%, without differences (F=73.3; P<0.0001) between these species, although there were differences with isolates Cv-3 and Cv-8 from C. muehlenbeckiae and isolates Cv-19 and Cv-29 of Curvularia sp., which caused an AFA of 10%. The isolates of C. spicifera (Cv-4) and C. muehlenbeckiae (Cv-3 and Cv-5) were the least virulent, since they caused an AFA of 3 to 9% (Table 3). In experiment 2, the percentages of AFA caused by isolates of C. ahvazensis (Cv-11) and Curvularia sp. (Cv-21) were 24.0 and 19.0%, respectively, with differences (F=27.0; P<0.0001) between these isolates, and the rest belonged to diverse species of Curvularia, which caused between 5.0 and 15.0% of LAA. In both experiments, the control plants sprayed with distilled water remained asymptomatic throughout the investigation (Figure 1B).

Figure 2 Phylogenetic tree of the glyceraldehyde-3-phosphate dehydrogenase (gapdh) gene, inferred by maximum likelihood (lnL -4286.90) in MEGA X. The scale represents the average of situations per site. Bootstrap values ≥70% are presented on the branches, near the internodes. The species found in this study are shown in bold, followed by a red circle. ATCC, American Type Culture Collection, Virginia, USA; BRIP, Queensland Plant Pathology Herbarium, Brisbane, Australia; CBS, Westerdijk Fungal Biodiversity Institute, Utrecht, the Netherlands; MFLUCC, Mae Fah Luang University Culture Collection, Chian Ria, Thailand; UTHSC, Fungus Testing Laboratory, Department of Pathology of the University of Texas Health Science Center, San Antonio, Texas, USA; CPC, Culture Collection of Pedro Crous, Westerdijk Fungal Biodiversity Institute; IMI, International Mycological Institute, CABI-Bioscience, Egham, Bakeham Lane, UK. ^ET , ^IsoT , ^IsoLT , ^SynT and ^T indicate ex-epitype, ex-isotype, ex-isolectotype, ex-syntype and ex-type strains, respectively. 

To comply with Koch’s standards, the inoculated Curvularia species were re-isolated from the symptomatic plants and their identity was confirmed by comparing the phenotypical characteristics of the cultures and morphometry of 20 conidia with the original isolates.

The species of Curvularia mentioned previously were proven to cause leaf spot in soybean plants, which contrasts with reports in Sinaloa in which Corynespora casiicola, Cercospora sojina and Cercospora kikushii are mentioned as causal agents of the disease (^{Rodríguez-Cota et al., 2019}), as well as in Brazil (^{Teramoto et al., 2017}), the United States (^{Sumabat et al., 2018}) and Argentina (^{Reznikov et al., 2018}). Although the symptoms caused by the Curvularia species in our study are similar to those caused by other fungi (^{Stewart and Rodríguez, 2016}; ^{Mwase and Kapooria, 2000}), these were not found in the sampling period of our investigation. Fungal foliar diseases in soybean that occur in the stage of grain filling are called late-season diseases (LSD), since they take place in plants in intermediate to advanced crop stages of maturity (^{De Lisi et al., 2015}). Given that foliar spot in soybean caused by C. spicifera, C. muehlenbeckiae, C. ahavenzis and Curvularia spp. in Sinaloa appears in the grain-filling stage, it must also be considered an LSD. This disease appeared in September and October, when rainfalls varied between 321 and 373 mm in the region and the temperature fluctuated between 26 and 39 °C (Junta Local del Valle del Carrizo weather station, Sinaloa, 2019). In other soybean-producing regions, Colletotrichum truncatum, Septoria glycines, Phomopsis sojae, Peronospora manshurica and Alternaria spp. have been registered as LSD causal agents (^{Carmona et al., 2015}), yet these pathogens were not found at the time the present investigation was being carried out. Studies in Argentina indicate that LSD reduce crop yield by 8 to 10%, and a maximum of up to 30% (^{Carmona et al., 2004}); in this sense, the investigations carried out in Sinaloa to determine the impact of the disease in crop yield and quality are justifiable, in the same way the implantation of strategies for their control would be, if needed.

Figure 3 A) Symptoms of foliar spot in leaflets of soybean plants from natural infections in the field; B) Non-inoculated control plant leaflets; C) Leaflets with symptoms, typical of leaf spot in soybean plants, caused by species of Curvularia (isolate Cv-11) in greenhouse inoculations. 

Conclusions

The fungi Curvularia spicifera, C. muehlenbeckiae, C. ahvazensis and Curvularia sp. were implicated in foliar spot in soybean, which began in the stage of grain filling (R6). The species were identified with the analysis of its morphometric characteristics and phylogenetic reconstruction, based on a fragment of the gene gapdh. The Curvularia species were found to be pathogenic in soybean var. Nainari and they displayed differen degrees of virulence. This is the first report in Mexico involving species of Curvularia in the etiology of leaf spot in soybean.