From 2022 to 2024, soybean samples with typical symptom of damping-off were collected from Jiangsu and Henan provinces of China (Figure 1). In total, 56 soybean samples displayed typical symptoms were collected from 11 counties in two provinces of China. The isolates were obtained by incubating the diseased samples at 20°C for 1-2 days in selective V8 juice agar (V8A) medium (15% clarified V8 juice with 2.5 g/L CaCO₃ and 2% agar) containing rifampicin, ampicillin, and pentachloronitrobenzene and then purifying them using single hyphal or colony tip culture methods. Oomycetes were cultivated on V8A, and the fungal strains isolated for our study were cultured on potato dextrose agar (PDA) medium at 25°C for at least 7 days before DNA extraction. The isolates used in this study are maintained in a collection at Jiangsu Vocational College of Agriculture and Forestry (JSAFC).

DNA was extracted from the pure cultures using the Universal Genomic DNA Extraction Kit (Solarbio, Beijing, China) following the manufacturer’s instructions. The ITS and Cox1 loci were amplified and sequenced with primers ITS5/ITS4 and OomCoxI-Levup/OomCoxI-Levlo, respectively (Robideau et al., 2011; White et al., 1990). PCR amplification and sequencing followed the protocol of Chen et al (Chen et al., 2017). Amplicon sequences were searched for similar sequences in National Center for Biotechnology Information (NCBI) using the basic local alignment search tool (BLASTn).

Phylogenetic reconstruction for the concatenated data set (ITS and Cox1) was performed utilizing maximum likelihood (ML) and Bayesian Inference (BI) methods in MEGA 11 (Tamura et al., 2021) and MrBayes 3.2.6 (Ronquist et al., 2012), respectively. Elongisporangium prolatum (CBS 845.68) was included as an outgroup. ML tree searches were performed assuming sequences evolved according to the GTR + G + I model and clade support was determined by 1000 bootstrap replicates. For BI analysis, four Markov chains were run for 30 million generations simultaneously, with samples taken from the posterior every 1000 generations. The first 25% of trees were discarded as the burn-in. Bootstrap proportions (BPs) and posterior probabilities (PPs) are used for branch values based on ML and BI, respectively.

The pure isolates of Globisporangium were cultured on V8A at 25°C for three days, and then 8-mm² mycelial plugs were cut from the edge of fresh colonies for pathogenicity tests. According with the inoculation method of Dorrance et al. (2008), hypocotyl slit inoculation technique with minor modifications has been used. Three-day-old soybean seedlings (cultivar Hefeng 47) were selected from germinating surface-sterilized seeds on WA plates. Each soybean hypocotyl was slightly scratched with a sterilized metal needle and placed a mycelial plug in contact with the wound. Each seedling was then transplanted into an individual pot containing 120 g sterilized soil that had been saturated with sterilized water. Each isolate was tested on three seedlings maintained at 25°C with a 12 h photoperiod in greenhouse conditions for 4-5 days. Three seedlings inoculated with sterile V8A plugs served as controls. Each isolate was tested on three seedlings per experimental run, and the entire experiment was independently replicated three times (total n=9 per isolate). After 4 days of inoculation, symptoms of infected seedlings were recorded. Disease severity index (DSI) was evaluated using the scale developed by Türkkan et al. (2022). Differences in severity among isolates was evaluated by the Tukey’s honestly significant difference test (P = 0.05) in SPSS 26.0 software (SPSS, Inc., Chicago, IL, USA). To fulfill Koch’s postulates, the infected seedlings were harvested for strain re-isolation on V8 agar medium, followed by ITS sequence comparison with original isolates.

The rpb1 (RNA Polymerase II Subunit B1) gene was selected as a candidate target for designing G. intermedium-specific LAMP primers based on bioinformatics analysis of the gene sequence among Globisporangium species. A total of ten LAMP primer sets were designed for G. intermedium using Primer Explorer ver. 5 (https://primerexplorer.eiken.co.jp/lampv5e/index.html), followed by a series of tests to assess both specificity and sensitivity (Feng et al., 2019c). Ultimately, an appropriate set was chosen for implementation in this investigation (Table 1). A forward inner primer (FIP) consisted of the complementary sequence of F1c and F2, a backward inner primer (BIP) consisted of B1c and B2, two outer primers (F3 and B3), and loop primer (LB) were used for LAMP, and the structure of the universal primers and their complementarity to target DNA is illustrated in Supplementary Figure S1.

The LAMP reaction was conducted in accordance with previously described methods (Chen et al., 2025; Dai et al., 2019; Feng et al., 2019a; Zhang et al., 2023). LAMP reaction mixtures with prereaction addition of Hydroxynaphthol blue (HNB) were optimized in preliminary tests by adjusting the concentration of reagents and incubation duration. The final LAMP reaction (26 μL volume) was performed by combining 1 μL Bst DNA polymerase (8 U·μL^-1, New England BioLabs, USA, Cat# M0275L), 3.5 μL dNTPs (10mM, TaKaRa Bio, Japan, Cat# 4030), 4 μL MgSO₄ (50 mM, Sigma-Aldrich, USA, Cat# M3409), 2.5 μL 10×ThermoPol Buffer (New England BioLabs, USA), 2 μL each of forward inner primer (FIP) and backward inner primer (BIP) (20 μM), 0.5 μL each of primers B3 and F3 (10 μM), 2 μL LB primers (20 μM), 4 μL betaine (5M, Sigma-Aldrich, USA, Cat# B0300), 2 μL HNB (2.4mM, Aladdin, USA, Cat# H196967), 2 μL DNA template, and sterilized double distilled water up to 26 μL. Each set of reactions included a control known to yield a positive detection result (100 ng of purified gDNA of G. intermedium as template) and a non-template control (NTC, DNA template replaced by double distilled water). The LAMP reaction mixtures were heated at a range of reaction temperatures (viz., 61, 62, 63, 64, and 65°C) for 63 min to select the optimal temperature (Supplementary Figure S2). Additionally, LAMP reactions were performed at the optimal reaction temperature (63°C) for 57, 60, 63, 66, 69 and 72 min to select the shortest viable reaction time (Supplementary Figure S2). The assays were evaluated by observation of the HNB color change from violet to blue, which denotes positive amplification, while a negative assay remains violet.

All primer sets were first checked for specificity using 100 ng of genomic DNA template from G. intermedium isolate JSAFC651. The isolate JSAFC651 was selected for sensitivity tests due to its robust growth and to challenge the assay under a potentially stringent condition. The established assay conditions were subsequently confirmed to be equally effective for all other G. intermedium isolates (Figure 2). One LAMP primer set with preliminarily determined specificity was further examined in a wide range of species. In total, 30 isolates of 15 Globisporangium species, three Phytopythium species, five Pythium species, two Phytophthora isolates, and five soil-borne fungal isolates were studied (Table 2). To evaluate the sensitivity of the method, a series of 10-fold dilutions from 10 ng/μL to 10 fg/μL of G. intermedium (JSAFC651) genomic DNA were utilized as templates, with NTC as the negative control. The LAMP assay was performed as previously described with the optimal LAMP primers and reaction parameters. The whole experiment was repeated twice to ensure reproducibility.

To evaluate the ability of LAMP to detect G. intermedium in infected plants, soybean cultivar “Hefeng 47” hypocotyls were inoculated with G. intermedium discs as mentioned above. Soybean hypocotyls treated with V8A plugs were served as the control group. Subsequently, the ‘Plant-LAMP’ (P-LAMP) detection methods, developed by Feng et al (Feng et al., 2015), has been utilized to analyses the pathogens in soybean tissues. Rotting or browned tissues were cut into 1 cm long (0.2 cm²) segments, collected in a 1.5-mL tube, mixed with 100 μL of sterile distilled water, vortexed for 60 s, and then 2 μL of the supernatant was used directly for LAMP amplification. In addition, the hypocotyl samples were placed on V8A and cultivated for 1-3 days at 25°C. The isolates for identification based on their ITS sequences. This experiment was carried out twice. Purified gDNA (100 ng) of G. intermedium isolate JSAFC651 was used as the template in a positive control reaction, while NTC served as the negative control.

Rhizosphere soil samples were collected from locations in Jiangsu and Henan provinces in China from 2022 to 2024, where G. intermedium was isolated from soybean roots, during preliminary surveys. A soil sample collected from the G. intermedium-free site at JSAFC was used as an additional negative control sample. All samples were transferred to JSAFC and stored at 4°C before further processing. The extraction of DNA from soil samples was efficiently accomplished utilizing a specialized kit known as the FastDNA Spin Kit for Soil (Solarbio, Beijing, China). The experiment was performed along with gDNA of G. intermedium (100 ng/μL) and NTC as positive and negative controls, respectively. The experiment was carried out twice.

The diseased samples were cultured in a selective medium, and 39 Pythium-like isolates were obtained. All isolates were initially identified via ITS sequencing, of which 23 showed at least 99.5% similarity to G. intermedium type specimen CBS 266.38. Subsequently, 9 of the 23 isolates were obtained from different area and further identified by polygenic phylogenetic analysis.

The analysis of the combined dataset (ITS + Cox1) of our isolates compared to reference sequences from ex-type or other authoritative specimens allowed us to allocate these isolates to molecular groups of known Globisporangium species (Table 3). The topological structure of the phylogenetic trees constructed using BI and ML criteria was basically consistent, demonstrating that the evolutionary relationships of the experimental strains were statistically supported. A consensus tree with clade support from BP and PP values was generated (Figure 3). The multigene phylogeny revealed that all nine Globisporangium isolates composed a strongly supported monophyletic clade (100% BP/1.00 PP) with the G. intermedium type specimen CBS 266.38, and were clearly distinguished from the other Globisporangium species. The nine G. intermedium isolates selected for phylogenetic analysis and assay validation originated from geographically distant locations (Jiangsu and Henan provinces) and exhibited unique haplotypes in the concatenated ITS-Cox1 sequence alignment, supporting their genetic distinctiveness and representative of the pathogen’s population in the surveyed regions.

The nine isolates selected from the polygenic phylogenetic analysis were further assessed for their pathogenicity. The result indicated that all G. intermedium isolates are pathogenic to soybean seedlings (Figure 4). When results of the mean disease severity were combined, G. intermedium isolate JSAFC651 exhibited higher virulence (mean DSl = 3.78) on soybean seedlings (P < 0.05). After four days of inoculation, soybean seedlings were dead after germination, or produced weaker seedlings with significantly shorter plant height, and shorter and fewer lateral roots, whereas those treated with sterile V8A plugs did not display any visible disease symptoms (Figure 4). In addition, G. intermedium could be reisolated from the symptomatic roots and was verified to be the original isolate based on comparisons of ITS sequences, thus fulfilling Koch’s postulates. These findings validated that G. intermedium is a soybean pathogen.

A total of ten primer sets targeting the rpb1 gene were designed according to the principles of species-specific LAMP primer design, of which showed the greatest specificity were further screened using a broad range of species (Table 2). The LAMP reaction mixtures containing gDNAs of nine G. intermedium isolates collected from Jiangsu and Henan provinces (Table 2) yielded positive detection as the color turned to blue (Figure 2). In contrast, LAMP reaction mixtures containing gDNAs of other oomycete or fungal species as well as NTC remained violet in color (Figure 2), indicating lack of detection of the target sequence. Two repeated LAMP assays yielded identical results against all tested 39 isolates. This result indicated that the primer set can specifically detect G. intermedium isolates. We further optimized the LAMP reactions for the primer set using temperature gradients from 61 to 65°C and reaction time gradients from 57 to 72 min. Considering the high specificity, reactivity, and stability, optimal temperature and time of 63°C and 63 min were selected respectively (Supplementary Figure S2).

Serial dilutions of genomic DNA from G. intermedium (JSAFC651) were employed to assess the detection limit of the LAMP method utilizing the chosen primer set at the optimal temperature. The concentrations of the DNA template spanned from 10 ng/μL to 10 fg/μL. As evidenced by a color alteration in the reaction products, the LAMP assay’s lowest detectable limit for G. intermedium was 10 pg/μL (Figure 5). Identical results were obtained from two replicate experiments.

Two days after inoculation with G. intermedium, typical symptoms emerged at the inoculation sites on the hypocotyls of the soybean cultivar “Hefeng 47”, while no such symptoms appeared on the non-inoculated controls. Crude DNA was then extracted from both the inoculated soybean seedlings and the non-inoculated ones. The LAMP reaction mixtures, incorporating DNA from soybean hypocotyls colonized by isolate JSAFC651, turned blue, paralleling the positive control that utilized G. intermedium gDNA as a template (Figure 6). In contrast, the non-inoculated samples retained their violet color. These consistent results across two independent assay repetitions highlight the LAMP assay’s high detection capability.

To evaluate the prevalence of detectable pathogens in field soil samples, we gathered a total of fifteen soil specimens from the vicinity of diseased soybean plants in Henan and Jiangsu provinces, China. When contrasted with the negative control group (one G. intermedium-free soil sample and one NTC), the LAMP assay successfully identified G. intermedium in the soil of all fifteen rhizosphere samples contaminated with G. intermedium, as well as in the positive control (Figure 7). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of GNP-LAMP were 100%, 100%, 100% (95% CI, 88.4 – 100.0) and 100% (95% CI, 39.8 – 100.0), respectively (Supplementary Table S1). The experiment was replicated twice, yielding consistent results in both runs.