Identification of Rice Seed-Derived Fusarium spp. and Development of LAMP Assay against Fusarium fujikuroi

Fusarium species are important seedborne pathogens that cause rice bakanae disease (RBD). In this study, 421 strains were isolated from 25 rice samples collected from Zhejiang, Anhui, and Jiangxi provinces of China. Furthermore, 407 isolates were identified as F. fujikuroi (80.05% isolation frequency), F. proliferatum (8.31%), F. equiseti (5.94%), F. incarnatum (2.61%), F. andiyazi (0.95%), and F. asiaticum (0.48%) based on morphology and translation elongation factor 1-alpha (TEF1-α) gene. Phylogenetic analysis of combined sequences of the RNA polymerase II largest subunit (RPB1), RNA polymerase II second largest subunit (RPB2), TEF1-α gene, and ribosomal DNA (rDNA) internal transcribed spacer (ITS) showed that 17 representative strains were attributed to six species. Pathogenicity tests showed that representative isolates possessed varying ability to cause symptoms of bakanae on rice seedlings. Moreover, the seed germination assay revealed that six isolates had different effects, such as inhibition of seed germination, as well as seed and bud rot. The loop mediated isothermal amplification (LAMP)-based assay were developed for the detection of F. fujikuroi. According to sequences of desaturase-coding gene promoter, a species-specific marker desM231 was developed for the detection of F. fujikuroi. The LAMP assay using seeds collected from field was validated, and diagnostics developed are efficient, rapid, and sensitive.


Introduction
Rice (Oryza sativa L.) is one of the most important staple foods that feeds more than 60% of the world's population [1]. Rice bakanae disease (RBD) caused by seed-borne Fusarium spp. may lead to a global food crisis. The disease symptoms of RBD are slender, chlorotic, and elongated primary leaves [2], due to the production of fungal gibberellic acid, which has resulted in the disease being called the Japanese word 'bakanae', which means the 'foolish seedling' [3]. The disease incidence usually ranges from 0.25% to 20% [4], even up to 40% in India [5], leading to a severe rice yield loss [6]. The pathogens responsible for the disease are primarily seedborne, but it also survives in soil and infected plant debris, which persists from one season to the next from the remains of infected plants [2]. Although species of the F. fujikuroi species complex (FFSC or fujikuroi) [7], such as F. andiyazi, F. fujikuroi, F. proliferatum, and F. verticillioides, are associated with bakanae of rice, F. fujikuroi is believed to be the core species responsible for the symptoms and was detected only in seed samples from Asia [8]. Besides, other species, like F. equiseti, F. graminearum, F. oxysporum, F. avenaceum, F. brevicatenulatum, F. begonia, F. sterilihyphosum, F. napiforme, and F. sporotrichioides, occasionally, also were isolated from seeds [4,9,10]. In

Seed Samples, Fungal Isolation, and Preliminary Identification
The rice seed samples, which were widely cultivated cultivars were collected from Zhejiang, Anhui, and Jiangxi provinces. A total of 25 seed samples, representing 23 varieties that were widely grown locally, were obtained (Supplementary Table S1). Rice seed samples for each variety was collected by a seed sampler using five bags (25 kg). After being mixed, one-hundred seeds of each sample were randomly selected [26], and they were placed into five plates (20 seed per plate) containing peptone-Pentachloronitrobenzene (PCNB) agar (PPA) medium (Peptone 15 g, KH 2 PO 4 1.0 g, MgSO 4 •7H 2 O 0.5 g, PCNB (Pentachloronitrobenzene) 750 mg, Agar 20 g, H 2 O to 1 L), which was a selective isolation medium for Fusarium spp. [27]; the plates were incubated at 25 • C for seven days. One isolate from per seed were selected, and it was moved into a plate containing potato dextrose agar (PDA, Guanzhi Biological Engineering Co. LTD, Shanghai, China) under ultraviolet (NUV) light for the single-spore isolation. The single-spore isolates were preliminarily identified based morphological characteristics on carnation leaf agar [28] and combination of sequencing of translation elongation factor 1α (TEF1-α) gene [9], as described below. The sequences of TEF1-α gene were analyzed by using the NCBI Basic Local Alignment Search Tool (BLAST) program in GenBank database.

Morphological Identification
To observe the morphological characteristics, the isolates were cultured on carnation leaf agar (CLA; 20 g/L bacto-agar, 5~10 carnation leaf pieces) [28] and PDA. The carnation leaf pieces were prepared from fresh carnation leaves. The leaves were cut into about 5 mm 2 pieces, dried in the air oven at 70 • C for 3~4 h, and then were sterilized using gamma irradiation from a Cobalt 60 source (Zhejiang University Testing Center, Hangzhou, Zhejiang, China). Fourteen days after culturing at 25 • C with NUV (near ultra violet), fungal asexual structures (the shape and size of microconidia, macroconidia, and conidiophores, length of microconidia chains, and sporodochia formation on carnation leaf pieces) were observed and measured using a Zeiss Axioskopt2 microscopy (Carl Zeiss Microscopy Company, Germany) with Axiocam CCD (charge coupled device) camera and Axiovision digital imaging software (AxioVision Software Release 3.1., v. ; Carl Zeiss Vision Imaging Systems)(Carl Zeiss Microimaging Company, Jena, Thuringen, Germany).

DNA Extraction, PCR Amplification, and Sequencing
All Fusarium isolates were cultured on potato dextrose broth (PDB) at 25 • C for five days. The fungal mycelia were harvested and grounded to a fine powder in liquid nitrogen. Genomic DNA was extracted by using a rapid fungal genomic DNA extraction kit (Sangon, Shanghai, China) according to the manufacturer's protocols. The genomic DNA obtained was resuspended in 50 µL Tris EDTA (ethylene diamine tetraacetic acid, TE) buffer and stored at −20 • C after the concentration of genomic DNA in each sample was quantified determined by a NanoDrop 2000 spectrophotometer (Thermo Scientific Inc., Wilmington, DE, USA).
To amplify sequences of different loci, the primer pairs EF1/EF2 for TEF1-α gene were used for PCR amplification of all isolates [29]. The representative isolates were selected for other PCR amplification according to geographical location, cultivation area of varieties and isolation frequency. The primer pairs RPB1-Fa/RPB1-G2R were used for PCR amplification for the RNA polymerase largest subunit gene (RPB1) [30], RPB2-5f2/RPB2-7cr for the RNA polymerase second largest subunit gene (RPB2) [30], and ITS (internal transcribed spacer) 6/ITS4 for amplification of ribosomal DNA (rDNA) ITS region. The PCR amplification was performed in a 25 µL reaction volume with 12.5 µL 2 × Taq Master (Sangon, Shanghai, China), 1.0 µL of each primer (10 µM), and 1 µL template DNA (2 µg/µL). The thermal cycling program was performed with 30 cycles after an initial denaturation at 95 • C for 3 min. Each cycle included a denaturation step at 95 • C for 1 min, annealing at a suitable temperature for 1 min, and an extension step at 72 • C for 2 min. Annealing temperatures for each reaction were 53 • C for ITS region, 55 • C for RPB1, RPB2, and 54 • C for TEF1-α. The amplified products were sent to the Hangzhou TsingKe Science and Technology Co., LTD (Hangzhou city, China), for sequencing in both directions.

Phylogenetic Analysis
The obtained sequences were analyzed by using BioEdit software [31], and they were deposited in the GenBank database (Table 2). Additional sequences mainly representing cultures from the Fusarium MLST (multilocus sequence typing) database (http://www. westerdijkinstitute.nl /fusarium/) and the NCBI's (National Center for Biotechnology Information) GenBank were downloaded and relevant sequences were included in the subsequent phylogenetic inference ( Table 2). The sequences of each gene were aligned with MAFFT (multiple alignment using fast Fourier transform, Kyoto University, Japan) 7.273 [32] and trimmed with BioEdit. The resulted alignment was put into the Gblocks 0.91 b to eliminate the ambiguously aligned positions and divergent regions prior to phylogenetic analyses [33].  The phylogenetic trees were constructed with maximum likelihood (ML) and Bayesian inference (BI) methods. The model of evolution applied to each alignment was estimated using jModel Test 2.1.7 [34] and the model chosen according to the Akaike information criterion. The best TIM1ef + I + G model was selected for RPB1, and CTR + I + G, TIM2 + G and JC + C for RPB1, TEF1-α and ITS, respectively. ML analyses were performed on the concatenated different sequences to produced different data sets, including RPB1 + RPB2 for genus Fusarium spp., RPB1+ RPB2 +TEF1-α for the FSSC (sambucinum) and FFSC (fujikuroi), and RPB2 + ITS + TEF1-α for FIESC (incarnatum-equiseti), with RaxmlGUI v. 1.5 [35]. ML bootstrap (ML-BS) analysis for each ML tree was performed with 1000 fast bootstrap replicates with the same parameter settings using the different model of nucleotide substitution for different data sets. A threshold of ≥80% was used as the cut-off for significantly supported nodes.
BI analyses were conducted with MrBayes v. 3.2.6 run by partitioning codon positions [36]. The best TrNef + I + G model was selected for RPB1 and RPB2, and TIM2ef + G and JC for TEF1-α and ITS. The Markov-chain Monte Carlo searches were performed with four chains, each of which was run for 10,000,000 generations, with trees sampled every 100 generations. The initial 25% of trees from each run were discarded as burn-in, and the remaining trees were combined into one tree with 50% majority rule consensus tree. BI posterior probability (BI-PP) values equal or above 0.95 were found to be significant. The resulted trees were imported into FigTree v.1.3.1. (http://tree.bio.ed.ac.uk/software/figtree/).

Pathogenicity Tests
The pathogenicity tests were performed to confirm the changes in virulence among different Fusarium species by using inoculated seeds of the rice cultivar (Oryza sativa L., cv. Jinzao 47). The rice seeds were sterilized by soaking in 1% sodium hypochlorite solution for 10 min, washed three times with sterile water, and then left to dry on sterile filter paper in a flow cabinet. For inoculum preparation, each single spore isolate, including F. fujikuroi ZJ01, F. andiyazi ZJ08, F. proliferatum ZJ05, F. incarnatum ZJ11, F. asiaticum ZJ10, and F. equiseti ZJ09, was cultured on PDA for 10 days; each plate was flooded with sterile distilled water, and spores were harvested by filtration with sterile miracloth. The spore suspensions were prepared and adjusted to 1 × 10 6 spores/mL using a hemocytometer, as described by Choi et al. [9]. The sterile seeds were soaked in the spore suspensions for 36 h or in sterile water as control. The eight inoculated seeds were sowed into a pot containing an autoclaved mixture of soil and sand in the ratio of 3:1. Each treatment had three replications for each isolate. The pots were placed in a greenhouse in which temperature was maintained at 28 • C during the day and 25 • C during the night. The whole test was repeated twice. The seedlings were observed daily after symptoms appeared.
The incidence of disease symptoms could start from the stage of seed germination due to the host-pathogen interactions. The seed germination assays were used to confirmed the effect six Fusarium species on seed germination. The sterilized seeds were placed into filter papers in a plate, 50 µL spore suspension (1 × 10 6 spores/mL) prepared above was added to each seed, and the seeds were dried in a laminar flow cabinet. The 20 inoculated seeds were transferred to a plastic tissue culture vessel (480 mL) containing three layers of wet filter paper, and the vessels were placed in incubators under a 12:12 h, light: dark, 28:25 • C regime. The control seeds were treated with sterilized water, and each treatment (20 seeds) was repeated five times for each isolate, and whole assay were repeated twice. The disease symptoms were observed daily and the percentage inhibition (PI) of seed germination was calculated using the formula PI (%) = [(C − N) /C)] × 100, where C was number of seed germination in control, and N was number of seed germination in different treatments by inoculation with fusarium isolates.

LAMP Primer Design
The dominant species associated with bakanae disease were selected as the target for detection based on identification results on rice seeds. If the dominant species was F. fujikuroi, sequences of enzyme-coding genes and their promoters in gibberellins biosynthesis pathway [17] were analyzed using genome data of Fusarium spp. in the GenBank database. If the conjectural target sequences were found, they were analyzed after PCR amplification and sequencing using the identified isolates. Once sequences of species-specific markers were obtained, LAMP primer design would be conducted using LAMP Primer Explorer V5 software (http://primerexplorer.jp/e/).

LAMP Reaction and Optimization of Its Condition
The LAMP reaction was performed in a 25 µL reaction mixture containing 3. Biolabs, UK). The reaction mixture was incubated at 50, 55, 60, 65, 67, 68, 69, and 70 • C, for 55 min to determine the optimal conditions for LAMP assay. For optimal time, the reaction mixture was incubated at 65 • C for 20,25,30,35,40,45,50,55, 60, 65, 70, and 75 min, respectively. Each reaction was terminated at 80 • C for 10 min. Each sample was amplified for LAMP assays with three independent replicates. The LAMP products were visualized by color change after adding 0.25 µL 10000× SYBR Green I (Sangon, Shanghai, China) and a short vortex or by 1% agarose gel electrophoresis.

Specificity and Sensitivity of LAMP Reaction
For testing stability of LAMP primers, 15 isolates of F. fujikuroi from different provinces were used for the detection of LAMP amplification. A total of 14 isolates of Fusarium. spp. and nine non-Fusarium species were used in this study to determine the specificity of LAMP primers (Supplementary Table S2). The genomic DNA for each isolate was extracted, as described above. The LAMP amplification for isolate were conducted at 65 • C for 60 min. The LAMP products were observed by a color change and product electrophoresis as described above. The PCR amplification of isolate were performed using the F3/B3 primer pair in an automated thermal cycler (Eppendorf AG, Germany), mainly including 54 • C annealing temperature and 30 cycles of denaturation in 25 µL reaction volume to confirm the reliability as described by Li [37]. The purified PCR products were verified by 1.0% agarose gel electrophoresis. The genomic DNA from F. fujikuroi was serially diluted from 9.4 ng/µL to 94 ag/µL to determine the LAMP assay's sensitivity. The LAMP amplification was performed as described above. PCR amplification was used as a control. The LAMP and PCR amplification for each concentration were repeated three times.

Detection of Rice Seed Samples
In order to determine the applicability of the LAMP assay on rice seeds, 30 seed samples were obtained. The rice seed samples were collected from infected (with typic symptoms of bakanae disease) or healthy rice panicles in 18 parcels of rice field with cv. Jinzao 47 during the harvest period in Fuyang city, Zhejiang province, and dried as general rice harvest procedure. After storage for 30 days, the genomic DNA of seeds was extracted using the Plant Direct PCR Kit (TsingKe Xinye biotechnology co., LTD, Beijing, China) according to the manufacturer's protocol. The LAMP amplification was performed as described above, and PCR amplification was used as control. LAMP and PCR amplification for each sample were repeated two times.

Isolation and Identification Using Morphology and TEF1-α Gene Sequences
For identification of Fusarium spp., variations of morphological characteristics from all isolates grown on CLA and PDA were compared based on their characteristics and divided into different species complex. Furthermore, the isolates were classified into the FFSC, when macroconidia produced by isolates were relatively slender with no significant curvature, and chlamydospores were absent. In this species complex, microconidia were abundant in false heads or shorter chains ( Figure 1A), and macroconidia were relatively slender, medium length with no significant curvature, 3-to 5-septate, apical cell tapered, and basal cell poorly developed ( Figure 1B), and they were identified as F. fujikuroi. The 337 isolates fell into this species. If microconidia were abundant, formed in chains and in false heads from polyphialides and monophialides ( Figure 1C), and were club-shaped, 0-septate ( Figure 1D); macroconidia were slender, relatively straight, 3-to 5-septate, and their apical cells were curved and basal cells poorly developed ( Figure 1E), and they were identified as F. proliferatum. The 35 isolates were attributed to this species. Similarly, if microconidia were clavate to oval with a flattened base, 0-septate and produced in abundance in long chains from monophialides ( Figure 1G), macroconidia were straight to slightly curved, and apical cell was slightly curved and basal cell was pedicillate, 3-to 6-septate, mostly 3-septate ( Figure 1F) and pseudochlamydospores were formed on this medium, they were identified as F. andiyazi. Four isolates belonged to this species, as described by Leslie [28] and Summerell [38]. Macroconidia produced by isolates were the pronounced dorsiventral curvature or falcate [20,28], and they were attributed to the FIESC. In this species complex, if the macroconidia from the sporodochia had pronounced dorsiventral curvature and their apical cells were more rounded comparing with F. compactum with the needle-like apical cells ( Figure 1H), and there were no microconidia on PDA, as described by Burgess [39], they were identified as F. equiseti. The 25 isolates were divided into this species. If the abundant conidia were produced from polyphialides in the aerial mycelia with "rabbit ears" appearance instead of in sporodochia ( Figure 1I), and they were straight, spindleshaped, with 3-4 septa ( Figure 1J), conidia also were known as mesoconidia [40]; 11 isolates producing mesoconidia were identified as F. incarnatum, which also was used as the more widely recognized name of F. semitectum [28]. Macroconidia produced by isolates were relatively slender, curved to almost straight, usually 5-septate, with a curved apical cell and a foot-shaped basal cell, and these isolates were attributed to the FSSC. In this species complex, if macroconidia were straight to slightly curved, 5-septate, apical cell slightly curved, and well-developed basal cell ( Figure 1K), very similar to that of F. graminearum, they were identified as F. asiaticum [37,41]. Two isolates were attributed to this species.
For confirming the results of morphological identification, 421 single-spore isolates were sequenced. Analysis of TEF1-α gene sequences revealed the 378 isolates had 99.9-100% identities with F. fujikuroi (337), F. asiaticum (2), F. proliferatum (35), and F. andiyazi (4), respectively, and they could clearly be distinguished by BLAST search, being identical to morphological identification. However, the 36 isolates had 99.9-100% similarity with F. equiseti or F. incarnatum, and apparently, two species were distinct, even though two species had a difference of significantly morphological characteristics ( Figure 1H-J) (Supplementary Table S3). Based on the sequence analysis, the other seven isolates with poor sporulation had 100% similarity with F. commune NT-LH01and identified as having F. commune (GenBank access No. MF150040.1). Among them, isolate ZJG 17 was deposited in GenBank sequence database (GenBank access No. MT560649).

Diversity of Fusarium spp. on Rice Seeds
Based on the results of identification using combination of morphological and molecular data, 414 isolates were initially divided into six Fusarium species (including F. fujikuroi, F. proliferatum, F. equiseti, F. asiaticum, F. incarnatum, and F. andiyazi), while another seven isolates were assigned to F. commune only based on TEF1-α gene sequences. According to results of isolation from three provinces (Supplementary Table S4), dominant species with the highest frequency were F. fujikuroi (80.05%), and the other species with low frequency in turn were F. proliferatum (8.31%), F. equiseti (5.94%), F. incarnatum (2.61%), F. commune (1.66%), F. andiyazi (0.95%), and F. asiaticum (0.48%). In comparison, seven isolates of F. commune were found only on cv. Xiushui 14 from Zhejiang region, and, obviously, they were not representative. The number and origin of these Fusarium spp. isolates are summarized in Supplementary Table S4.

Phylogenetic Studies
According to geographical location and cultivation area of varieties, 18 representative isolates were selected for PCR amplification (Table 2). Approximately 1500 bp were amplified for RPB1 gene, 900 bp for RPB2, 730 bp for TEF1-α, and 500 bp for ITS region. The first analysis based on RPB1 and RPB2 sequences was conducted to evaluate the preliminary identification of the isolates and their phylogenetic affinity among the different species complexes of Fusarium. For ML analyses, a GTR + I + G model was selected for all two gene regions and incorporated into the analyses. The combined RPB1 and RPB2 sequences dataset included 43 in group taxa with 18 test isolates and F. torreyae 54149 as outgroup taxon. The phylogenetic analyses showed that the 18 test isolates were distributed in three Fusarium species complexes, namely the FFSC (fujikuroi, 13 isolates), FIESC (incarnatum-equiseti, three isolates), and FSSC (sambucinum, two isolate) ( Figure 2). To describe the species boundaries of FFSC, phylogenetic tree was conducted by using the three-locus combined RPB1, RPB2, and TEF1-α sequence dataset with 26 taxa. The phylogenetic analyses showed that the three test isolates ZJ05, ZJ06, and AH03 and reference F. proliferatum 2294 clustered as a clade, forming a sister clade with reference F. proliferatum 43617 with higher ML-BS (100%) and high BI-PP (1.00) support (Figure 3). The eight test isolates (ZJ 01-04, JX01-02, and AH01-02) clustered with reference F. fujikuroi 13566 as a clade, forming a sister clade with reference F. fujikuroi 43610. Similarly, the two test isolates ZJ07 and ZJ08 clustered with reference F. andiyazi 119857 as a clade (Figure 3).   Similarly, for analyses of the FSSC, phylogenetic analysis was conducted by using the three-locus combined RPB1, RPB2, and TEF1-α sequence dataset with 10 taxa, including two isolates test ZJ10 and ZJ12 taxa, and F. dimerum 36140 was used as outgroup. Phylogenetic analysis showed that the two test isolates ZJ10 and ZJ12 clustered with F. asiaticum 13818 as a distinct clade with higher ML-BS (100%) and high BI-PP (1.00) support ( Figure 5).

Pathogenicity of Fusarium spp.
The isolates from six species were used for pathogenicity tests with inoculated seeds to evaluate the virulence differences between different Fusarium species. After fifteen days of sowing, symptoms of bakanae disease was observed on seedlings inoculated with F. fujikuroi ZJ01, while individual yellowing leaves appeared on seedlings inoculated with F. andiyazi ZJ08 and F. proliferatum ZJ05, and relative stunted plants with yellowing leaves were observed on seedlings inoculated with F. incarnatum ZJ11, F. asiaticum ZJ10, and F. equiseti ZJ09. However, no symptoms were observed on sterile water-inoculated seedling controls. After twenty days of sowing, typical bakanae disease symptoms were observed on seedlings inoculated with F. fujikuroi ZJ01 (Figure 6), as described by Choi [9]. The yellowing leaves, including few dying leaf/leaves, were observed on normal height seedlings inoculated with F. andiyazi ZJ08 and F. proliferatum ZJ05. In comparison, the yellowing leaves, including few dying leaves, Fwere visible on stunted seedlings inoculated with F. incarnatum ZJ11, F. asiaticum ZJ10, and F. equiseti ZJ09. The results revealed that isolates of six species were all pathogens and able to cause disease, but there were differences in virulence. In vitro seed germination tests were conducted using inoculated seeds to reveal the reason for differences in virulence among Fusarium species. The test results showed that the isolates of six Fusarium species all influenced germination of rice seeds, including complete inhibition of seed germination (non-germinated), seed and sprouts decay, as well as seedling stunt ten days after inoculation (Supplementary Figure S3). Inoculation increased percentage inhibition (PI) of seeds germination by 73% in F. asiaticum ZJ10, 71% in F. andiyazi ZJ08, 54% in F. incarnatum ZJ11, 44% in F. equiseti ZJ09, and 35% in F. proliferatum ZJ05, respectively, but there was no significant difference between F. asiaticum ZJ10, and F. andiyazi ZJ08, as well as among F. incarnatum ZJ11, F. proliferatum ZJ05, and F. equiseti ZJ09 (p < 0.05). The repeated tests also showed that the seed germination PI value was almost the same as the above result (Supplementary Table S5). In addition, observation of symptoms indicated that small dark spots with the size of a pinpoint appeared first at the top of a sprout three days after inoculation. Subsequently, clear sprout decay symptoms were observed at 5-7 days, as shown in Supplementary Figure S3. After ten days, F. andiyazi ZJ08 inoculation showed the highest percentage of sprouts decay (PSD) by 68.8 ± 8.5%, followed by PSD by 62.5 ± 8.7% in F. asiaticum ZJ10, 50.0 ± 4.1% in F. incarnatum ZJ11, 37.5 ± 9.6% in F. equiseti ZJ09, 28.8 ± 6.3% in F. proliferatum ZJ05, and 27.5 ± 6.5% in F. fujikuroi ZJ01, respectively. Still, there was no significant difference between F. andiyazi ZJ08 and F. asiaticum and among F. fujikuroi ZJ01, F. proliferatum ZJ05, and F. equiseti ZJ09 (p < 0.05) (Supplementary Table S5). Although the inoculation with F. proliferatum ZJ05 gave the greatest percentage of non-germinated seeds (PUS) by 17.5 ± 4.3%, followed by PUS by 16.3 ± 6.5% in F. equiseti ZJ09, 8.8 ± 6.5% in F. asiaticum ZJ10, 8.8 ± 5.4% in F. andiyazi ZJ08, 6.3 ± 5.4% in F. fujikuroi ZJ01, and 2.5 ± 2.5% in F. incarnatum ZJ11, respectively, there was no significant difference was observed between F. proliferatum ZJ05 and F. equiseti ZJ09 and among F. andiyazi ZJ08, F. asiaticum, F. fujikuroi ZJ01, and F. incarnatum ZJ11 (p < 0.05) (Supplementary Table S5).

LAMP Primers
The promoters and coding sequences analysis of enzyme-coding genes in gibberellins biosynthesis pathway in Fusarium spp. showed that the promoter sequences of desaturase gene (des) in F. fujikuroi (Supplementary sequence 1) had 81.26% identity with F. proliferatum (GenBank access No. AJ628021) instead of no similarity with other Fusarium. spp. in the GenBank database. Therefore, according to tis conserved region of the promoter sequences (Supplementary sequence 2), the primer pair DesF/DesR was designed, and they were used for PCR amplification from 23 isolates of Fusarium spp. and other plant fungal pathogens (Supplementary Table S2).
The PCR products were observed only in the isolates of F. fujikuroi, F. proliferatum, and F. oxysporum. The sequence analysis showed that the promoter sequences from F. fujikuroi had a relatively higher similarity with F. proliferatum than F. oxysporum. A 231 bp sequence was selected from promoter region of the desaturase-coding gene of F. fujikuroi as the marker (Supplementary Figure S6), which was designated as DesM 231 , and LAMP primers were designed by the Primer Explorer V5 software based on DesM 231 sequence. The location and direction of primers are indicated in the Supplementary Figure S6. The primer sequences are shown in Table 3.

LAMP Reaction and Optimization of Its Conditions
The LAMP assay was carried out using F. fujikuroi DNA as a template at 50, 55, 60, 65, 67, 68, 69, and 70 • C, respectively, for 55 min. Although LAMP products were observed from 55 to 68 • C for 55 min according to color change (visual inspection) and electrophoresis on agarose, the best results were obtained when the reaction temperature was maintained at 65 • C (Supplementary Figure S7A,B).
Furthermore, the LMAP test was carried out at 65 • C for different times, including 25,30,35,40,45,50,55,60,65,70, and 75 min. The results showed that LAMP products turned to green color in tubes treated from 25 to 75 min, and only tubers used as control and treated for 20 min turned to orange (Supplementary Figure S8A); however, the best results were produced after reaction time was maintained for 60 min based on electrophoresis (Supplementary Figure S8B).

Specificity and Sensitivity of LAMP Reaction
The LAMP assay was conducted using the15 isolates of F. fujikuroi at 65 • C for 60 min, and amplification products were observed by visual inspection in all 15 isolates ( Figure 7A), showing reliability of primers. To confirm LAMP assay specificity, the 23 isolates of Fusarium spp. and other plant fungal pathogens (Supplementary Table S2) were used as the LAMP assay, which was conducted at 65 • C for 60 min. The results demonstrated that LAMP products turned to green color only in F. fujikuroi ZJ01 and turned to orange in other isolates ( Figure 7B), showing that LAMP primers were species-specific. Meanwhile, the results of PCR amplification also displayed that PCR product was observed only in F. fujikuroi ZJ01 ( Figure 7C), further confirming the results of the LAMP assay. For verifying the sensitivity of LAMP assay, genomic DNA extracted from F. Fujikuroi ZJ01 with concentration of 9.4 ng, 0.94 ng, 94 pg, 9.4 pg, 0.94 pg, 94 fg, 9.4 fg, 0.94 fg, 94 ag, respectively, was used as DNA template for LAMP and PCR amplification at 65 • C for 60 min. The results of LAMP amplification revealed that about 9.4 fg/µL template was sufficient for LAMP-based diagnostics ( Figure 8A,B). In comparison, approximate 94 pg/µL template was the lowest concentration for conventional PCR detection ( Figure 8C) and showed that this detection limit was 10,000-times lower than that of LAMP amplification.

Detection of Rice Seed Samples
The LAMP assay was used for the detection of F. fujikuroi in natural seeds. The results of LAMP amplification showed that LAMP products amplifying from 18 samples of diseased rice seeds all turned to green color ( Figure 9A), while, from 12 samples of healthy rice seeds, all turned to orange, displaying the reliability of LAMP reaction. Similarly, the PCR amplification results using DNA template of the same concentration for each sample demonstrated that PCR products were found only in 18 samples of diseased but no PCR product in 12 samples of healthy rice seed samples ( Figure 9B).

Diversity of Fusarium spp. on Rice Seeds
This is the first study to systematically analyze the Fusarium species on rice seed from three provinces of China. The identification of the Fusarium species revealed that 89.31% of the isolates belong to the FFSC, showing them to be dominant that is similar to previous studies [4,9,42]. Among the FFSC, although four species (F. andiyazi, F. fujikuroi, F. proliferatum and F. verticillioides) were deemed to be associated with RBD [8], F. verticillioides was not found in this study. In addition, another four species with low isolation frequency in this study, belong to FIESC (F. incarnatum and F. equiseti), FSSC (F. asiaticum), and F. oxysporum species complex (FOSC; F. commune) [43], were also found on rice seeds in other countries [4,9,20]. However, F. commune was found in only one variety that was isolated in the group.
The isolation frequency and species variation on rice seeds depend on the origin (different countries, regions, and varieties), showing diversity of Fusarium spp. However, differences in species composition between studies may be due to various factors, e.g., sample size, target organ (plant or seed), geographical distribution, and climate effect of species distribution [44]. In China, rice is one of the prominent cereal crops, and about 65% of Chinese people rely on rice. Although this survey of the Fusarium species has its limitations in geographical range and number of varieties in rice-growing regions of China, our results provide an important information for nationwide survey of diversity of the Fusarium species in future.

Identification of Fusarium spp.
The identification of Fusarium spp. associated with bakanae disease only depending on morphological characteristics is a highly challenging task due to its diverse characteristics [9]. Thus, standardized culture media and methods, along with multi-locus molecular data, are currently required for confident identification at the species level. Although some studies have demonstrated that the FFSC associated with bakanae, including other several species (F. thapsinum, F. pseudonygamai, F. moniliforme, and F. commune), can be clearly distinguished by only the TEF1-α gene sequence [8,9,41], no evidence showed that TEF1-α gene was suitable to delimitation of all Fusarium species. However, a combination of RPB1 and RPB2 genes were used to infer evolutionary relationships of Fusarium species complex [7], while combination of RPB1, RPB2 and TEF1-α genes even were used for genetic diversity analysis of Fusarium oxysporum f. sp. Cubense [45] and especially recommended [30]. However, due to lacking available reference sequences of RPB1 gene in F. incarnatum, six species identified by morphology and TEF1-α gene in this study, wherefore, were first divided into the species complex only using RPB1 and RPB2 genes due to lacking references sequences of individual species (Figure 2), and then each species complex was analyzed further using different loci sequences. In the FFSC, morphologically microconidia of F. fujikuroi forms are oval or club-shaped with a flattened base having 0 to 1 septum, and false heads and chains of short to medium length are produced from poly-and monophialides [28]. The macroconidia are relatively slender with tapering apical, and poorly developed basal cell with 3 to 5 septa and chlamydospores are absent. Especially, F. fujikuroi has very similar morphological characteristics with F. proliferatum. Phylogenetically, F. fujikuroi and F. proliferatum are closely related species. In the FIESC, species belonging to the incarnatum or equiseti morphotype morphologically have been described by Avila [20], although morphological markers cannot distinguish most of the cryptic species in the FIESC. In the FSSC, F. andiyazi is morphologically very similar to F. verticillioides, and the difference is that F. andiyazi produces pseudo-chlamydospores only in this species [28], and their close relationship also was confirmed by phylogenetic analysis, as showed in Figure 2. In the FSSC, F. asiaticum is morphologically indistinguishable from F. graminearum, and they have slightly different conidial features [40], but phylogenetic analysis supports their species delimitation, as showed in Figure 2. Our results confirmed that RPB1, RPB2, and TEF1-α genes are reliable molecular markers similar to other studies [30,44]. This will provide an important information for identification of Fusarium spp. on rice seed.
In addition, due to the fact that the relationship of each SC (species complex) with corresponding secondary metabolites has been established by Reference [7,20], mycotosins, like trichothecene in the FSSC and FIESC and fumonisins in FFSC, could be predicted, which provided available clues for analyzing pathogenicity of Fusarium spp.

Pathogenicity of Fusarium spp.
F. fujikuroi isolates were able to cause bakanae disease, and other species also have been studied [2,3]. However, various studies have found that the FFSC (including F. proliferatum, F. verticillioides, and F. andiyazi) were associated with the bakanae disease on rice [8,9,46,47]. In this study, isolate of six species were capable of causing symptoms inherent to bakanae, but the difference of symptoms was significant ( Figure 6). The isolate of F. fujikuroi promoted the elongation of seedlings, isolate of F. andiyazi, and F. proliferatum caused the relative stunting of seedlings comparing with control, being consistent with other reported findings [8,41]. Interestingly, the severe stunting occurred in seedlings inoculated with isolates of F. equiseti, F. asiaticum, and F. incarnatum, respectively. Although F. concentricum is capable of causing symptoms of RBD [48], and F. incarnatum (F. semitectum) is a pathogen of bakanae [49], F. equiseti is known as a saprophytic colonizer [9], while the role of F. asiaticum as a saprophytic or pathogenic fungus has not been documented. In vitro tests of seed germination were conducted to reveal the reason for different symptoms in pathogenicity. The results showed that six species isolate all influenced rice seed germination (Supplementary Figure S3). The isolates of F. asiaticum and F. andiyazi caused the highest percentage inhibition of seed germination, followed by F. equiseti, F. incarnatum, F. proliferatum, and F. fujikuroi (Supplementary Figure S4). The inhibitory mechanism of seed germination is related to mycotoxins.

Application of LAMP as Diagnostic Tool
The LAMP method possesses the advantage of fast, sensitive, reliable, and inexpensive technology over PCR and is widely used as diagnostic tool. However, sensitivity of assay involves the design of suitable primers and the optimization of reaction conditions, while specificity of LAMP assay depends on the uniqueness of the target marker. In detection of F. fujikuroi, for example, although the TEF1 gene [23] and intergenic spacer (IGS) region [36] are used as markers, successful LAMP amplification requires relatively strict reaction conditions due to high similarity of marker sequences. In comparison, sequences of unique or very different fragments are used for LAMP primer design, and reliable results are obtained easily. In this study, the promoter sequence of desaturase in Gibberellins biosynthesis pathway was used as the target marker (DesM231) due to difference of larger bases between F. fujikuroi and other species of Fusarium (Supplementary Figure S6). LAMP amplification was observed under various conditions, as described above, and the sensitivity of LAMP reaction was increased for more than 9.4 fg under an optimization condition, and assays for detecting rice seeds from the fields further confirm that the developed LAMP technique using for detection of F. fujikuroi is reliable. In addition, LAMP amplification using DesM231 marker provides information related to GA3 production (from its precursors GA4 and GA7) (Supplementary Figure S5). Although the non-ribosomal peptide synthetase gene (NRPS31) in F. fujikuroi, consisting of 11-gene cluster (apicidin synthetase, APS) without production of apicidin [56], is used as a marker in LAMP assays [6,25], NRPS31 is closely related to the NRPS APS1 (66-71% identical) in F. incarnatum, which is part of a 12-gene cluster responsible for synthesis of apicidin, a histone deacetylase inhibitor with antiparasitic activity [57,58], showing that it is not unique in genus Fusarium. Furthermore, a previous study showed that F. fujikuroi m567 (Genbank accession no. AJ417493) contained all of the GA-biosynthesis genes and produced GA3 [18]. In comparison, its mutant analysis revealed that GA4 to GA7 and GA7 to GA3 were converted by the 13-hydroxylase, respectively. In this study, although GA3 production is not detected, pathogenicity tests provide indirect evidence and sequences of desaturase promoters of F. fujikuroi having 99.88% identities with F. fujikuroi m567. Therefore, GA3 production could be anticipative in F. fujikuroi isolates. This indicates that LAMP-based detection developed and standardized for the detection of F. fujikuroi associated with RBD on seeds is a powerful tool for the rapid and sensitive detection of F. fujikuroi and predicting GA3 production.
Supplementary Materials: The following are available online at https://www.mdpi.com/2076-081 7/10/1/1/s1. Figure S1: Geographical location of Zhejiang, Anhui and Jiangxi provinces in China. Red color: Zhejiang province; Yellow: Anhui province; blue: Jiangxi province. Figure S2: Typical symptoms of bakanae on rice seedlings in field. A. Abnormal elongation plants with chlorotic leaves. B. A partial magnification in Figure A. Figure S3: Inhibitory effect of different Fusarium spp. on rice seed germination 10 days after inoculation. CK: Control; FF: F. fujikuroi ZJ01; FAN: F. andiyazi ZJ08; FP: F. proliferatum ZJ05; FI: F. incarnatum ZJ11; FAS: F. asiaticum ZJ10; FE: F. equiseti ZJ09. Arrowheads: inhibition of seed germination; arrows: seed rot; double arrows: bud rot. Figure S4: Percentage inhibition of seed germination after inoculation with isolates of Fusarium fujikuroi ZJ01, F. proliferatum ZJ05, F. andiyazi ZJ08, F. asiaticum ZJ10, F. incarnatum ZJ11 and F. equiseti ZJ09 for 10 days. The same letter(s) are not different significantly at p < 0.05 according to LSD. Figure S5: Substrate and product of desaturase marked in red in the last step of gibberellin biosynthesis pathways. Figure S6: Partial sequences of the desaturase gene promoters of in Fusarium fujikuroi ZJ01and F. proliferatum ZJ05 and location of Loop-Mediated Isothermal Amplification (LAMP) primers. Arrows indicate LAMP primers (inner and outer primers), along with their position for detection of F. fujikuroi. Figure S7 Table  S1: Rice varieties and origin in this study. Table S2: Strains of fungi used for testing the specificity in the loop mediates isothermal amplification (LAMP) and polymerase chain reaction (PCR) assays. Table S3: GenBank accessions No. of 402 Fusarium isolates used for phylogenetic analysis in this study. Table S4: Number and isolation frequency of Fusarium species from 25 rice varieties in three provinces. Table S5: The results of rice seed germination tests by inoculation with six Fusarium spp. Isolates. Sequence1: Desaturase Gene from NBCI database. Sequence2: Partial sequences of desaturase promoter region amplified using primer pair DesF/DesR.

Data Availability Statement:
The data presented in this study are available in insert article or supplementary material here.