Isolation of Main Pathogens Causing Postharvest Disease in Fresh Codonopsis pilosula during Different Storage Stages and Ozone Control against Disease and Mycotoxin Accumulation

Codonopsis pilosula is an important Chinese herbal medicine. However, fresh C. pilosula is prone to decay during storage due to microorganism infections, seriously affecting the medicinal value and even causing mycotoxin accumulation. Therefore, it is necessary to study the pathogens present and develop efficient control strategies to mitigate their detrimental effects on the herbs during storage. In this study, fresh C. pilosula was collected from Min County in Gansu Province, China. The natural disease symptoms were observed during different storage stages, and the pathogens causing C. pilosula postharvest decay were isolated from the infected fresh C. pilosula. Morphological and molecular identification were performed, and pathogenicity was tested using Koch’s postulates. In addition, the control of ozone was examined against the isolates and mycotoxin accumulation. The results indicated that the naturally occurring symptom increased progressively with the extension of storage time. The mucor rot caused by Mucor was first observed on day 7, followed by root rot caused by Fusarium on day 14. Blue mold disease caused by Penicillum expansum was detected as the most serious postharvest disease on day 28. Pink rot disease caused by Trichothecium roseum was observed on day 56. Moreover, ozone treatment significantly decreased the development of postharvest disease and inhibited the accumulations of patulin, deoxynivalenol, 15-Acetyl-deoxynivalenol, and HT-2 toxin.

Owing to its unique climate, Gansu Province in China has more than 2000 years of history in cultivating C. pilosula. With its high commercial value and profitability, the cultivation of C. pilosula is continuously expanding. However, some soil-borne fungi, bacteria, and nematodes have adapted to the local host plants and natural environment in the region, thus, resulting in serious diseases that severely affect the yield and quality of C. pilosula.
Research to date has focused on field disease of C. pilosula. For instance, Zhao et al. [8] suggested that Fusarium oxysporum is the predominant pathogen causing root rot of C. pilosula in Dingxi County, Gansu Province. Using molecular technology, Yu et al. [9] isolated and identified five pathogens in China: Puccinia Campanumoeae Pat causing rust disease, Helicobasidium mompa Tanaka causing violet root rot, Sphaerotheca Codonopisi (Golov.) Z.Y. Zhao causing powdery mildew, Fusarium oxysporum Schl. causing root rot, and Septoria codonopsidis Ziling causing blight of Codonopsis tangshen in Chongqing. Based on morphology and molecular characteristics, Wang et al. [10] isolated and identified Septoria codonopsidis Ziling causing leaf spot of C. pilosula in Gansu Province, while Chen et al. [11] proposed that Botrytis cinerea is the typical fungus causing gray mold in C. pilosula.
Modern research indicates that fresh Chinese herbs have higher active ingredients and pharmacological activities than those of dry products [12]. For instance, the contents of flavonoids, saponins, and polysaccharides in fresh Astragalus were 1.5 times more than those in dry products [13]. Because of this, the market for freshly harvested Chinese herbs has grown rapidly and is today more significant and well known than before. Unfortunately, the postharvest losses due to the disease of fresh Chinese herbs are quite severe. Nevertheless, no reports are available regarding the prevalence of pathogens during this period. In general, the herb is usually harvested in late autumn and allowed to dry naturally on the field for more than two months before being transported to the traders. The freshly harvested C. pilosula must be sufficiently dried (about 16% of water content), then stored for 3-6 years. If C. pilosula does not reach an optimum level of dry matter, its abundance storage of fat, starch, protein, sugar, and other organic substances may favor the development of some latent pathogens and ultimately result in serious postharvest disease. Fungal infections are thought to result in yearly losses of 15% to 25% [14]. Such postharvest losses cause enormous economic damage to C. pilosula processing industries, and more importantly, the product may completely lose its medicinal value, becoming contaminated with mycotoxins that have carcinogenic, teratogenic, and mutagenic toxicity [15]. It is therefore important to systematically study the postharvest diseases of freshly harvested C. pilosula, identify the pathogens present at various storage stages, and develop an efficient control strategy to mitigate their effects.
Ozone, an antioxidant compound, has been widely applied to manage the postharvest decay of fruits and vegetables. The efficacy of ozone in controlling postharvest disease is mainly ascribed to its strong inhibitory activity on pathogenic fungi. However, there are no reports on the influence of ozone on fresh Chinese herb medicine postharvest disease and mycotoxin production.
In this study, we collected fresh C. pilosula from Min County in Gansu Province, China, investigated the disease development of freshly harvested C. pilosula during storage, then isolated and identified pathogens causing C. pilosula postharvest disease based on morphological and molecular biological techniques during different storage stages. Finally, we examined the influence of ozone on fresh C. pilosula postharvest disease and mycotoxin production.

Sample
Samples of freshly harvested C. pilosula (cv. Mindang) were obtained from Min County (location 35 • N and 104 • E) in Gansu Province, China. Sample roots of similar size and without obvious pest or mechanical damage were selected in October 2020, and transported to the Chemical Biology Laboratory, College of Science, Gansu Agricultural University within 24 h after bagging, and stored at room temperature for further analysis.

Disease Development of Freshly Harvested C. pilosula during Different Storage Stages
Freshly harvested C. pilosula samples (without any processing treatment such as washing and drying) were placed directly in plastic bags (20 • C, 50 % RH) in darkness for 7, 14, 21, 28, 42, and 56 days. Subsequently, the naturally occurring symptoms were observed and described. Different pathogens cause different disease symptoms, for example, in the early stages of root rot, small brown spots appear on the surface of the lower fibrous or lateral roots, with mild decay. As the disease expands, it gradually spreads to the main roots. The roots gradually decay from the bottom upwards and become dark brown and waterlogged [9]. Three replicates were included in each treatment, each treatment consisted of 50 samples of C. pilosulas, with a total of 900 samples being included in the whole experiment (50 samples × 3 replicates × 6 time points).

Isolation and Purification of Pathogens Causing Disease of Freshly Harvested C. pilosula during Different Storage Stages
Pathogens were isolated and purified from C. pilosula based on typical decay symptoms (the appearance of mycelium and spores on the surface of C. pilosula) [16]. Fragments (5 mm × 5 mm) showing typical disease symptoms during different storage stages were excised with sterilized scalpels from the junction of the healthy and diseased tissue, and the fragments were disinfected using 1% NaClO for 3 min, and then rinsed with sterile water three times to remove any NaClO residue. The treated fragments were placed onto potato dextrose agar (PDA) medium and cultured in darkness at 25 • C for 5 to 7 d. Single colonies were picked and transferred to a new PDA medium with a puncher. After 4 to 5 cycles of purification, a single purified colony was obtained. The preliminary identification was conducted based on colony morphology and macro and microconidia characteristics [17]. The spore suspension (1 × 10 6 spores/mL) was inoculated on a PDA plate, and after drying, a cover glass was inserted into the PDA medium at an angle of 45 • , and then cultured at 25 • C for 2 to 5 d. Subsequently, to observe colony morphology and pigment production, a 2 µL spore suspension was inoculated centrally in the PDA medium at 25 • C for 7 to 9 d. Colony morphology and pigment production were recorded on day 9. Mycelia were transferred using the copper pick-in method [18]. All samples were sprayed gold with an ion sputtering apparatus (MSP-1S, Shenzhen Research Precision Instrument Co., Ltd., Shenzhen, China) under 220 V and 40 mA, then the morphology of the spores was observed using a scanning electron microscope (SEM) (JSM-5910LV, Japanese electronics company, Tokyo, Japan).
The PCR amplification procedure was as follows: pre-denaturation at 94 • C for 5 min, denaturation at 94 • C for 10 s, annealing at 53 • C for 10 s, extension at 72 • C for 30 s, 3 cycles, and holding at 72 • C for 5 min. The amplified products were subject to electrophoresis with 2% agarose gel. The amplified fragments were sequenced by Beijing Bomede Biotechnology Co., Ltd., and the sequences were aligned to the NCBI (https://www.ncbi.nlm.nih.gov/ (accessed on 8 July 2021)) using BLAST for homology analysis. A phylogenetic tree was constructed by MEGA7.0 software (Molecular Evolutionary Genomics Analysis Version 7) using the neighbor-joining method.

Pathogenicity Test
The healthy and freshly harvested C. pilosula was thoroughly washed with tap water to remove any adhering soil and air dried. The herbs were surface disinfected with 0.1% NaClO for 15 min, rinsed with sterile water three times to remove excessive NaClO, and air dried again. Spore suspensions (1 × 10 6 spores/mL) of the above isolated pathogens were prepared [18] and artificially inoculated by spraying onto the surface of the fresh and healthy C. pilosula. The control group was inoculated with sterile water. After natural drying, the inoculated C. pilosula were kept in darkness (20 • C, 50% RH). After an incubation period of 28 days, the disease symptoms were recorded, and different pathogens led to different disease symptoms [23]. The pathogens were re-isolated again from the C. pilosula's infected roots and stem, and their similarity to the original isolates' morphological characteristics was verified. The casual isolates that obeyed the criteria specified by Koch's postulates were conducted for further study. Healthy C. pilosula were treated and inoculated according to the above Section 2.5. Sterile water spraying inoculation was regarded as control. Gaseous ozone was generated by the OSAN ozone generator (Aoshan Huanbao Technology Industry Co., Ltd., Dalian, China). The concentration of ozone (2 mg L −1 ) was adjusted using an ozone detector [18]. Ozone fumigation was performed in a closed transparent bag (80 cm long × 60 cm wide) (25 • C, relative humidity 75%), and the inoculated samples were subjected to ozone treatment for 1 and 2 h each day, respectively. The treatment continued for 7 days. Subsequently, the treated tissue was stored for 56 days in plastic bags (22 ± 2 • C, 75-80%). The disease development was evaluated by statistically determining the disease index and natural incidence according to the report by Sha et al. [24] with minor modifications (Table 1). Each treatment contained three replicates, and one replicate included 50 samples. The diseased area accounts for 1-5% of the total area of C. pilosula 2 The diseased area accounts for 6-25% of the total area of C. pilosula 3 The diseased area accounts for 25-50% of the total area of C. pilosula 4 The diseased area accounts for 51-75% of the total area of C. pilosula 5 The diseased area accounts for 76-100% of the total area of C. pilosula Class frequency: the number of diseased plants at each rate; Score of rating class: the diseased value for each rate.

Effect of Ozone Treatment on the Mycotoxin Accumulation in the Rotten Tissue
For mycotoxins analysis, the samples treated by ozone fumigation were collected and the rotten tissue was excised from the diseased root and immediately kept in liquid nitrogen and stored at −80 • C until mycotoxin analysis. Each treatment contained three replicates, and one replicate included 50 samples.
A 5.0 g frozen sample was ground in liquid nitrogen, then was transferred to a 50 mL centrifuge tube with extraction solvent to extract mycotoxin. For different kinds of mycotoxin, various purification and detection methods were employed, patulin (PAT) purification and detection was carried out as described by the method [25]; trichothecene purification and detection was performed according to the method [26].

Statistical Analysis
The data of disease index, disease incidence, and mycotoxin accumulation were expressed as the means [standard error (±SE)] using Duncan's multiple range tests. The experiment of the effect of ozone on postharvest disease and mycotoxin accumulation of C. pilosula was performed at least three times. Statistical analyses were performed using SPSS v.17.0 (SPSS, Inc., Chicago, IL, USA), and Duncan's multiple range test (p < 0.05) was employed in this study.

Disease Development of Freshly Harvested C. pilosula during Different Storage Stages
With the extension of storage time, the disease development of freshly harvested C. pilosula was more severe (Figure 1). When stored for 7 days, mild symptoms of disease were observed, and a small number of white hyphae and moldy spores were found on the main root and lateral root of C. pilosula. When stored for 14 days, hyphae were gradually diffused, and more white hyphae and moldy spores covered the root of C. pilosula. After 21 days of storage, colonies expanded and their color changed slightly from white to light green, with vigorous mycelium growth. When stored for 28 days, some yellow, red, and green hyphae appeared on the surface of the C. pilosula. When stored for 42 days, multiple colonies were distributed over the surface of C. pilosula and the tissue was damaged. After 56 days of storage, C. pilosula was seriously diseased, and the tissue was wrinkled, soft, and even rotten.
cates, and one replicate included 50 samples.
A 5.0 g frozen sample was ground in liquid nitrogen, then was transferred to a 50 mL centrifuge tube with extraction solvent to extract mycotoxin. For different kinds of mycotoxin, various purification and detection methods were employed, patulin (PAT) purification and detection was carried out as described by the method [25]; trichothecene purification and detection was performed according to the method [26].

Statistical Analysis
The data of disease index, disease incidence, and mycotoxin accumulation were expressed as the means [standard error (±SE)] using Duncan's multiple range tests. The experiment of the effect of ozone on postharvest disease and mycotoxin accumulation of C. pilosula was performed at least three times. Statistical analyses were performed using SPSS v.17.0 (SPSS, Inc., Chicago, IL, USA), and Duncan's multiple range test (p < 0.05) was employed in this study.

Disease Development of Freshly Harvested C. pilosula during Different Storage Stages
With the extension of storage time, the disease development of freshly harvested C. pilosula was more severe (Figure 1). When stored for 7 days, mild symptoms of disease were observed, and a small number of white hyphae and moldy spores were found on the main root and lateral root of C. pilosula. When stored for 14 days, hyphae were gradually diffused, and more white hyphae and moldy spores covered the root of C. pilosula. After 21 days of storage, colonies expanded and their color changed slightly from white to light green, with vigorous mycelium growth. When stored for 28 days, some yellow, red, and green hyphae appeared on the surface of the C. pilosula. When stored for 42 days, multiple colonies were distributed over the surface of C. pilosula and the tissue was damaged. After 56 days of storage, C. pilosula was seriously diseased, and the tissue was wrinkled, soft, and even rotten.

The Isolation of Pathogen from C. pilosula with Postharvest Disease during Different Storage Stages
During the whole storage period, a total of nine isolates of pathogenic fungi were isolated and purified. The nine different fungi were obtained using repeated plate streaking during different storage stages. For example, on the 7th day of storage, 2 isolates were obtained, which were named 7-1 and 7-2, respectively; on the 14th day of storage, 5 isolates were obtained, among them, 3 isolates were newly obtained, which were named 14-1, 14-2, and 14-3, respectively; on the 21st day of storage, 6 isolates were obtained, among them, a new isolate was named 21-1; on the 28th day of storage, 8 isolates were obtained, and two of them were newly isolated, named 28-1 and 28-2. There were no new isolates on the 42nd and 49th day. On the 56th day of storage, 9 isolates were obtained, and there was a new isolate named 56-1.

Morphological Identification of Pathogens at Different Storage Stages
After isolation and purification, the pathogens were cultivated on a PDA medium. Colony morphology, spore morphology, and sporangiophore morphology of the nine isolates of pathogenic fungi were observed.
For isolate 7-1, the mycelia grew fast at a rate of 24.57 mm/d on the PDA plate. The colony texture was flocculent with a white color, irregular edge, and no pigment ( Figure 2A). Spores were spherical or nearly spherical in size at 4.8-5.6 µm × 5.0-6.2 µm ( Figure 3A); sporangia were spherical with a diameter of 78-81 µm ( Figure 4A). For isolate 7-2, the mycelia grew fast at a rate of 21.5 mm/d on the PDA plate. The colony texture was flocculent and white or gray-white with dense mycelia, a neat edge, and no pigment ( Figure 2B); spores were spherical or subspherical and approximately 3.9-4.5 µm × 4.3-5.9 µm in size ( Figure 3B); sporangia located at the end of hyphae were spherical, with a diameter of 59-63 µm ( Figure 4B) ( Table 2).
lates were obtained, among them, 3 isolates were newly obtained, which were named 14-1, 14-2, and 14-3, respectively; on the 21st day of storage, 6 isolates were obtained, among them, a new isolate was named 21-1; on the 28th day of storage, 8 isolates were obtained, and two of them were newly isolated, named 28-1 and 28-2. There were no new isolates on the 42nd and 49th day. On the 56th day of storage, 9 isolates were obtained, and there was a new isolate named 56-1.

Morphological Identification of Pathogens at Different Storage Stages
After isolation and purification, the pathogens were cultivated on a PDA medium. Colony morphology, spore morphology, and sporangiophore morphology of the nine isolates of pathogenic fungi were observed.
For isolate 7-1, the mycelia grew fast at a rate of 24.57 mm/d on the PDA plate. The colony texture was flocculent with a white color, irregular edge, and no pigment ( Figure  2A). Spores were spherical or nearly spherical in size at 4.8-5.6 μm × 5.0-6.2 μm ( Figure  3A); sporangia were spherical with a diameter of 78-81 μm ( Figure 4A). For isolate 7-2, the mycelia grew fast at a rate of 21.5 mm/d on the PDA plate. The colony texture was flocculent and white or gray-white with dense mycelia, a neat edge, and no pigment (Figure 2B); spores were spherical or subspherical and approximately 3.9-4.5 μm × 4.3-5.9 μm in size ( Figure 3B); sporangia located at the end of hyphae were spherical, with a diameter of 59-63 μm ( Figure 4B) ( Table 2).  For isolate 14-1, the mycelia grew at a rate of 7.39 mm/d on the PDA plate. The colony was round or nearly round with a fluffy texture, neat or wavy edges, and a thick and dense mycelium; the secreted pigment was cinnamon colored, the surface and back of the edge were rose-red ( Figure 2C). The number of conidia was small, the shape was oval or fusiform with a size of 2.4-3.0 µm × 3.0-3.6 µm ( Figure 3C); 2 to 4 septa were observed, and the conidiophores had branches ( Figure 4C). For isolate 14-2, the mycelia grew at a speed of 4.75 mm/d on the PDA plate. The colony texture was villous or cotton floc, with a light pink color, and the colony edge was light pink or white, with dense mycelia in the middle of the colony and sparse mycelia at the border of the colony ( Figure 2D); conidia were oblong or fusiform and 2.1-3.2 µm × 2.8-3.5 µm in size ( Figure 3D); conidiophores were erect and branched ( Figure 4D). For isolate 14-3, the mycelia grew at a speed of 4.57 mm/d on the PDA plate. The colony was round with neat edges, pale purple or dark purple in color, and the edges were white or grayish-white with flocculent or fluffy hyphae ( Figure 2E). The conidia with 3 to 4 septa were elongated oval, oval, or slightly curved, measuring 2.2-3.3 µm × 2.8-3.5 µm ( Figure 3E); conidiophores had branches ( Figure 4E).    For isolate 21-1, the mycelia grew at a speed of 4.29 mm/d on the PDA plate. The colony's front color was white, while the back was pale yellow; mycelia were creeping and looser with a neat edge ( Figure 2F); conidia were spherical or nearly spherical with a size of 2.3-3.2 µm × 2.8-3.5 µm (Figure 3F), and the conidiophores had branches ( Figure 4F).
For isolate 28-1, the mycelia grew at a rate of 8.71 mm/d on the PDA plate. The front color of the colony was gray-green with dense granular texture, and the reverse was white ( Figure 2G); conidia were striate, with colorless monospores, globose or oblate, with a size of 2.1-3.4 µm × 3.4-4.2 µm ( Figure 3G); conidiophores were erect, septate, colorless, apex branched, and broom shaped ( Figure 4G). For isolate 28-2, the mycelia grew with a speed of 6.50 mm/d on the PDA plate. A blue-green, and yellowish-brown with white areas were observed at the front and back of the colony, respectively. The center of the colony had protrusions and was slightly flocculent at the central surface and the texture was fluffy and powdery, forming several granular concentric rings ( Figure 2H); Conidia were striate, spherical or ellipsoidal, with a size of 2.4-2.8 µm × 2.8-3.6 µm ( Figure 3H); conidiophores were erect, septate, colorless, apex branched, and broom shaped ( Figure 4H).
For isolate 56-1, the mycelia grew with a speed of 10.14 mm/d on the PDA plate. An orange-pink, and pink with annual rings was observed on the front and back of the colony ( Figure 2I); the conidia were loosely gathered at the top of the mycelium, and were obovate or pear shaped, colorless, bicellular, with a size of 5.8-7.0 µm × 10-14 µm ( Figure 3I); conidiophores were erect, colorless, with an ultimate swelling, and 2.0-3.5 µm × 100-160 µm in size ( Figure 4I).

Pathogenicity Test
Pathogenicity tests were used to verify the pathogenicity of the isolates causing postharvest disease according to Koch's postulates. During the whole incubation, for C. pilosula inoculated with 7-1 (Actinomucor elegans), dense white flocculent hyphae covered the main root on the 7th day, grew rapidly with time, and the expansion of colonies on the 28th day led to the decay of the epidermis of C. pilosula and leakage of sap ( Figure 7A). For C. pilosula inoculated with 7-2 (Mucor hiemalis), gray flocculent hyphae appeared on the 7th day, grew rapidly on the 14th day, covering the whole tissue rotted on the 28th day ( Figure 7B). For C. pilosula inoculated with 14-1 (Fusarium acuminatum): white dense mycelia were observed on the 7th day, the color of the dense mycelia gradually became rose-red on the 14th day; a rose-red glue was secreted and leaked from the C. pilosula tissue on 28th day ( Figure 7C). For C. pilosula inoculated with 14-2 (Fusarium equiseti), there was less white velvet mycelial growth in the early stage, the colony became larger on the 14th day, and the color of C. pilosula epidermis turned pale pink or light yellow on the 28th day ( Figure 7D). For C. pilosula inoculated with 14-3 (Fusarium oxysporum), there were white velvet colonies on the main root and lateral root on the 7th day, the color became light purple on the 14th day, and the color of C. pilosula epidermis turned dark purple on 28th day ( Figure 7E). For C. pilosula inoculated with 21-1 (Clonostachys rosea), small white villous colonies were found on the surface of the tissue on the 7th day, the colony gradually expanded on the 14th day, and turned yellow and the tissue of C. pilosula produced yellowish glue on the 28th day ( Figure 7F); For C. pilosula inoculated with 28-1 (Penicillium expansum), granular mold spots (2 to 3 mm in diameter) were observed on the surface of C. pilosula at the initial stage, a cluster of colonies gradually formed on the 14th day with a layer of blue powder on the surface; the plaque spread continuously, and the color of C. pilosula epidermis became gray and green, the tissue at the lesion developed soft rot and mildew on the 28th day ( Figure 7G). For C. pilosula inoculated with 28-2 (Penicillium aurantiogriseum), small white granular mold spots appeared on the 7th day; the plaque expanded, and the number of colonies increased on the 14th day; the color of colonies turned blue and green, and C. pilosula developed serious mildew on the 28th day ( Figure 7H).
For C. pilosula inoculated with 56-1 (Trichothecium roseum), white plaque with a diameter of 2 to 3 mm was initially observed on the 7th day; the orange-pink plaque expanded irregularly on the 14th day; the tissue was wrinkled with soft collapses and a dark color on the 28th day ( Figure 7I). C. pilosula inoculated with the nine isolates had typical symptoms that were similar to the original natural symptoms. Based on the above typical and similar symptoms, the nine isolates were again identified through isolation, purification, and cultivation on PDA culture, and the same morphological and molecular biological characteristics were observed.
C. pilosula at the initial stage, a cluster of colonies gradually formed on the 14th day with a layer of blue powder on the surface; the plaque spread continuously, and the color of C. pilosula epidermis became gray and green, the tissue at the lesion developed soft rot and mildew on the 28 th day ( Figure 7G). For C. pilosula inoculated with 28-2 (Penicillium aurantiogriseum), small white granular mold spots appeared on the 7th day; the plaque expanded, and the number of colonies increased on the 14 th day; the color of colonies turned blue and green, and C. pilosula developed serious mildew on the 28th day ( Figure 7H).

Ozone Fumigation Inhibited the Development of C. pilosula Postharvest Disease
The development of postharvest disease was effectively suppressed in C. pilosula inoculated with nine isolates after ozone fumigation treatment, and there was an ozoneexposure-time-dependent relationship with the inhibitory effect. For instance, the disease indexes in C. pilosula inoculated with Actinomucor elegans (7-1) after 1 and 2 h of ozone exposure were, respectively, 70% and 37% higher than those in control ( Figure 8A). The disease incidences in C. pilosula inoculated with F. acuminatum (14-1) after 1 and 2 h ozone exposure were, respectively, 69% and 36% higher than those in control ( Figure 8B). Similar results were obtained in the other isolates inoculated with C. pilosula after 1 and 2 h ozone exposure for 56 days of storage ( Figure 8). It was obvious that with prolonged ozone treatment, the disease index and disease incidence dropped sharply. For C. pilosula inoculated with 56-1 (Trichothecium roseum), white plaque with a diameter of 2 to 3 mm was initially observed on the 7th day; the orange-pink plaque expanded irregularly on the 14th day; the tissue was wrinkled with soft collapses and a dark color on the 28th day ( Figure 7I). C. pilosula inoculated with the nine isolates had typical symptoms that were similar to the original natural symptoms. Based on the above typical and similar symptoms, the nine isolates were again identified through isolation, purification, and cultivation on PDA culture, and the same morphological and molecular biological characteristics were observed.

Ozone Fumigation Inhibited the Development of C. pilosula Postharvest Disease
The development of postharvest disease was effectively suppressed in C. pilosula inoculated with nine isolates after ozone fumigation treatment, and there was an ozoneexposure-time-dependent relationship with the inhibitory effect. For instance, the disease indexes in C. pilosula inoculated with Actinomucor elegans (7-1) after 1 and 2 h of ozone exposure were, respectively, 70% and 37% higher than those in control ( Figure 8A). The disease incidences in C. pilosula inoculated with F. acuminatum (14-1) after 1 and 2 h ozone exposure were, respectively, 69% and 36% higher than those in control ( Figure 8B). Similar results were obtained in the other isolates inoculated with C. pilosula after 1 and 2 h ozone exposure for 56 days of storage ( Figure 8). It was obvious that with prolonged ozone treatment, the disease index and disease incidence dropped sharply.

Ozone Fumigation Inhibited the Mycotoxin Accumulation in the Rotten Tissue
More importantly, ozone treatment significantly inhibited mycotoxin accumulation in the rotten tissue of the inoculated C. pilosula. For instance, the content of patulin (PAT) of the rotten tissue in C. pilosula inoculated with P. expansum (28-1) was significantly (p < 0.05) inhibited by 38.9 % and 53.0 %, respectively, after 1 and 2 h of ozone exposure ( Figure  9B). Similarly, the concentrations of 15-ADON and HT-2 of the rotten tissue in C. pilosula inoculated with F. acuminatum (14-1) were markedly decreased by 35.1 % and 59.9 % (15ADON), respectively, and 33.2 % and 50.8 (HT-2), respectively, after 1 and 2 h ozone exposure ( Figure 9A). Similar results were also found in C. pilosula inoculated with T. roseum (56-1) after ozone application ( Figure 9C). For the control group and other patho-

Ozone Fumigation Inhibited the Mycotoxin Accumulation in the Rotten Tissue
More importantly, ozone treatment significantly inhibited mycotoxin accumulation in the rotten tissue of the inoculated C. pilosula. For instance, the content of patulin (PAT) of the rotten tissue in C. pilosula inoculated with P. expansum (28-1) was significantly (p < 0.05) inhibited by 38.9% and 53.0%, respectively, after 1 and 2 h of ozone exposure ( Figure 9B). Similarly, the concentrations of 15-ADON and HT-2 of the rotten tissue in C. pilosula inoculated with F. acuminatum (14-1) were markedly decreased by 35.1% and 59.9% (15ADON), respectively, and 33.2 % and 50.8 (HT-2), respectively, after 1 and 2 h ozone exposure ( Figure 9A). Similar results were also found in C. pilosula inoculated with T. roseum (56-1) after ozone application ( Figure 9C). For the control group and other pathogens infected with fresh C. pilosula, no mycotoxins were found (Figure 9).

Discussion
At present, numerous studies have addressed the pathogens causing preharvest disease in Chinese herbs, these reports indicated that Fusarium species are the dominant pathogens that cause preharvest disease in the various regions [28][29][30][31]. However, there are limited reports on postharvest diseases of freshly harvested Chinese herbal medicines during storage. Of relevance, Chen et al. [32] isolated and identified P. crustosum, P. viridicatum, P. aurantiogriseum, and P. brevicompactum from Angelica sinensis and C. pilosula decoction pieces during storage. Nevertheless, there are significant differences between Chinese medicinal decoction pieces and freshly harvested Chinese herbs as experimental materials. Chinese medicinal decoction pieces are mostly obtained from the Chinese herbal medicine market, and these products are available in dry or dehydrated forms. On the other hand, freshly harvested Chinese herbs come from the field, without any processing and treatments after harvest such as drying or dehydration. Therefore, fresh C. pilosula is more susceptible to molds and decay, owing to the higher water content and the abundance of nourishing substances that allow the growth of pathogens.
To our knowledge, this is the first report to isolate and characterize the pathogens causing postharvest disease of C. pilosula during different storage periods. The results indicated that, with the extension of storage time, the symptoms of the disease were more severe. Moreover, a total of 9 isolates were isolated and characterized based on the mor-

Discussion
At present, numerous studies have addressed the pathogens causing preharvest disease in Chinese herbs, these reports indicated that Fusarium species are the dominant pathogens that cause preharvest disease in the various regions [28][29][30][31]. However, there are limited reports on postharvest diseases of freshly harvested Chinese herbal medicines during storage. Of relevance, Chen et al. [32] isolated and identified P. crustosum, P. viridicatum, P. aurantiogriseum, and P. brevicompactum from Angelica sinensis and C. pilosula decoction pieces during storage. Nevertheless, there are significant differences between Chinese medicinal decoction pieces and freshly harvested Chinese herbs as experimental materials. Chinese medicinal decoction pieces are mostly obtained from the Chinese herbal medicine market, and these products are available in dry or dehydrated forms. On the other hand, freshly harvested Chinese herbs come from the field, without any processing and treatments after harvest such as drying or dehydration. Therefore, fresh C. pilosula is more susceptible to molds and decay, owing to the higher water content and the abundance of nourishing substances that allow the growth of pathogens.
To our knowledge, this is the first report to isolate and characterize the pathogens causing postharvest disease of C. pilosula during different storage periods. The results indicated that, with the extension of storage time, the symptoms of the disease were more severe. Moreover, a total of 9 isolates were isolated and characterized based on the morphology and biological characters during storage for 56 days. All the nine isolates could cause different postharvest disease by Koch's postulate. However, the results went beyond Koch's postulate because Koch's postulate is mainly applied to one pathogen. However, we initially observed in the postharvest disease that many pathogens were interacting for disease, and we isolated and identified the different pathogens, then inoculated the healthy C. pilosula with the obtained pathogen. Finally, we observed the different disease symptoms caused by different pathogens. Therefore, we thought Koch's postulate had limitations for the results of this study.
A. elegans and M. hiemalis were the main pathogens causing mucor rot of freshly harvested C. pilosula on the 7th day of storage. After 14 days of storage, Fusarium species of F. acuminatum, F. equiseti, and F. oxysporum were the major pathogenic fungi causing root rot in C. pilosula. After 21 days of storage, besides the above-isolated pathogens, a new pathogen of C. rosea was isolated and characterized. On the 28th day of storage, blue mold caused by Penicillum spp. was the typical postharvest disease, and two new pathogens of P. expansum and P. aurantiogriseum were isolated and identified. On the 56th day of storage, a new pathogen T. roseum was isolated and characterized.
From the above results, the first postharvest disease was caused by Mucor infection on the 7th day of storage, and the main pathogens were A. elegans and M. hiemalis. A. elegans grew rapidly on PDA plates, and it took around 3 days to cover the whole petri dish. The morphology of colonies, spores and sporangiophores were similar to those isolated from necrotic skin lesions in humans, and the predominant fungal pathogen causing invasive mucormycosis for humans [33]. In addition, A. elegans, which can result in sufu's white flake, was discovered in the sufu [34]. Mucor hiemalis had a fast growth speed on PDA medium, with a white colony initial color, which changed gradually to grayish brown ( Figure 3B). The spore sacs were ellipsoid, and sporangiophores were erect and branched ( Figure 4B), This observation was similar to the morphology of M. hiemalis causing mucor rot of mandarin fruit in California reported by Saito et al. [35]. The spores of the two Mucor species can survive in the air and infect hosts mainly through the air.
During storage of 14 days, Fusarium spp. were the main pathogens causing postharvest root rot of C. pilosula; among them, F. acuminatum had the strongest pathogenicity, and also produced mycotoxins, with nearly round colonies, neat edges, vigorous mycelia, flocculent, and dense ( Figure 2C). Most conidia were fusiform and septate ( Figure 3C), whose morphology was similar to F. acuminatum from garlic bulb rot in Serbia [36]. Moreover, F. acuminatum also metabolizes 15ADON and HT-2 toxins (Figure 9). The toxins 15-ADON and HT-2 are attributed to trichothecene with detrimental effects, including cytotoxicity, acute toxicity, immunotoxicity, and chronic toxicity, and pose a serious threat to human health. Tan et al. [37] suggested that DON and 3-ADON were detected in Fusarium head blight (FHB) of wheat infected by F. acuminatum. The colony of F. equiseti was white, and the mycelium had cotton-like flocculence ( Figure 2D). Conidia were oval ( Figure 3D), and the morphology was consistent with the observation by Afroz et al. [22], who mentioned that F. equiseti isolated from cabbage Fusarium wilt in Korea had a morphology of loosely floccose and whitish-brown aerial mycelia, and the pigmentation of pale orange on PDA medium; moreover, the size and morphology of macroconidia and chlamydospores were also in accordance with our observation. The F. oxysporum colony was initially white, then turned off-white later, to light purple-dark purple pigment. Hyphae were flocculent or villous and denser ( Figure 2E), the conidia were elongated oval and slightly pointed at both ends ( Figure 3E), which was consistent with the morphology of F. oxysporum isolated from sesame plants. Previous study revealed that Fusarium species is attributed to "latent infected pathogen" [38], the infection of Fusarium usually occur in the field (growth stage of plants) during the blooming or heading of plants. The spores adhere to the surface of the leaf, then transported into the calyx, finally colonize on the root of plants, and remain in a latent status until there are favorable environmental conditions (such as high temperature and humidity) [39]. Therefore, it is possible to manage plant disease more effectively during the pathogen infection time.
After 21 days of storage, C. rosea was isolated and identified, but its disease infection was not severe. The hyphae were white with neat edges, mycelium was creeping and loose ( Figure 2F), and the morphology was similar to that isolated from root rot of Astragalus membranaceus in China [40]. C. rosea has been regarded as a typical biocontrol fungus to inhibit pathogenic fungal growth. The most serious postharvest disease was blue mold caused by Penicillium spp., including P. expansum and P. aurantiogriseum. P. expansum is a typical postharvest pathogen of pome fruits (such as apples and pears) causing blue mold and patulin contamination. The isolated P. expansum colony was grayish-green, and the edge was white, with a granular or velvet dense texture ( Figure 2G). The microconidia were colorless with monospores that were globose or oblate. The number of conidia was high, and the conidiophores were erect, separated, and colorless ( Figure 3G), whose morphology was consistent with the blue mold of apple fruits as reported by Bahri et al. [41]. In the present study, patulin (PAT) was detected in C. pilosula infected by P. expansum. Patulin is a secondary metabolite generated by Aspergillus and Penicillium species under favorable conditions, and usually found in blue mold disease of pome fruits and their corresponding products. The maximum limit of PAT in fruit juice was set as 50 µg/kg by the European Union [42]; however, no standard of PAT level limits is established in Chinese herbal medicines. P. aurantiogriseum, a kind of plant endophytic fungi, is also a casual pathogen causing fruit postharvest decay. Shim et al. [43] isolated the pathogen from pear fruit infected with blue mold in Korea, and Liu et al. [44] isolated P. aurantiogriseum from fresh-cut lettuce corruption. The cultivation characters and microscopic examination ( Figures 2H, 3H and 4H) in our study were very similar to the above reports' morphological characteristics. Penicillium species are attributed to "wound pathogens" [38] that mainly infect host plants through wounds sustained by fruit cracking or mechanical damage during harvesting, transportation, handling, and storage. Therefore, to avoid pathogen infection, delicate operations should be recommended during packing and handling.
T. roseum was identified on the 56th day of storage, as the predominant fungus causing C. pilosula disease. T. roseum is a typical necuotrophic fungal pathogen that can infect various postharvest fruits and vegetables, leading to trichothecenes contamination. Sharma et al. [45] isolated T. roseum from postharvest pink rot of avocado in Israel. T. roseum is also known to cause pink rot in muskmelon and apple fruit. The cultivation characters and microscopic examination of the pathogen from the pink rot of fruit were consistent with our observation. Moreover, 15-ADON, DON, and HT-2 toxins were found in the rotten tissue of C. pilosula pink rot caused by T. roseum, which were trichothecenes. Tang et al. [46] also suggested T-2 toxin and neosolaniol (NEO) were detected in the core rot of apple fruit infected by T. roseum. T. roseum is attributed to "wound pathogen", which mainly infects host plants through wounds; however, it can also carry out latent infection when the plant grows in the field, then end its colonization and development, without showing any symptoms until the plant is harvested. Therefore, it is a challenging job to manage the disease caused by T. roseum. The combination application of preharvest spray and postharvest treatment is proposed.
Ozone, a strong oxidant compound, is widely used to control postharvest decay. In the present study, ozone treatment significantly reduced the development of postharvest disease of freshly harvested C. pilosula, especially, when the herbs were exposed to 2 h treatment compared to 1 h. The result was in accordance with the report that ozone exposure for 10 min was more effective than 5 min in controlling pear disease [47]. The inhibitory effect of ozone was attributed to oxidation. Ozone can attack the cell plasma of plant pathogenic fungus, leading to membrane lipid peroxidation, and cell membrane integrity destruction, and loss of pathogenicity to the plant host [18]. In addition, we found that, as the ozone exposure time increased, the control effect became more visible. The reason is maybe that there is a cumulative effect for ozone exposure, the cumulative effect of 2 h ozone exposure was more than that of 1 h ozone exposure.
More importantly, we found that ozone treatment greatly suppressed the mycotoxin accumulation in the rotten tissue of the infected C. pilosula (Figure 9). A similar result was documented by Xue et al. [48], who indicated ozone application significantly reduced NEO accumulation. Considering the two reasons, ozone treatment controlled postharvest disease by inhibiting fungi growth; on the other hand, ozone can act directly with the chemical structure of mycotoxin, thus destroying the structure of mycotoxin [48].

Conclusions
A total of nine isolates were identified and characterized by morphology and molecular technology from postharvest diseases of C. pilosula during different storage stages. The nine isolates, with varying morphologies of colony, spore, and conidiophore, could cause several symptomatic postharvest diseases during different storage stages. Postharvest diseases of C. pilosula during storage can also metabolize and produce mycotoxins, which pose health threats to humans. Ozone application not only controlled postharvest diseases of C. pilosula during storage but also reduced mycotoxin accumulation. Therefore, in order to effectively control postharvest decay, with different pathogens having various infection modes and infection periods, the incorporation application of preharvest spray and postharvest treatment should be proposed; on the other hand, ozone application needs to be widely recommended and the mechanism of ozone action on disease and mycotoxins suppression should be further studied.