DNA Metabarcoding Reveals the Fungal Community on the Surface of Lonicerae Japonicae Flos, an Edible and Medicinal Herb

Lonicerae Japonicae Flos (LJF) has been globally applied as an herbal medicine and tea. A number of reports recently revealed fungal and mycotoxin contamination in medicinal herbs. It is essential to analyze the fungal community in LJF to provide an early warning for supervision. In this study, the fungal community in LJF samples was identified through DNA metabarcoding. A total of 18 LJF samples were collected and divided based on the collection areas and processing methods. The results indicated that Ascomycota was the dominant phylum. At the genus level, Rhizopus was the most abundant, followed by Erysiphe and Fusarium. Ten pathogenic fungi were detected among the 41 identified species. Moreover, Rhizopus, Fusarium, and Aspergillus had lower relative abundances in LJF samples under oven drying than under other processing methods. This work is expected to provide comprehensive knowledge of the fungal community in LJF and a theoretical reference for enhanced processing methods in practical manufacturing.


Introduction
Lonicerae Japonicae Flos (LJF), a traditional Chinese medicine, is dried from flower buds or newly boomed flowers of Lonicera japonica Thunb.[1][2][3].LJF was first recorded in Shen Nong Ben Cao Jing for its medicinal effect on alleviating fever and detoxification in the East Han Dynasty [2].In modern pharmacology, LJF was first recorded in the Pharmacopoeia of the People's Republic of China in 1995, and now, it has been used in more than 500 prescription drugs [4,5].According to recent reports, LJF has played an important role in the production of anticancer and anti-COVID-19 drugs globally [6,7].In addition, LJF is considered an herbal tea for its function of clearing away heat and toxins, and it has received considerable popularity for a long time in East Asia [8].Traditionally, LJF mainly grows in Shandong, Hebei, and Henan provinces in China [3,4,9,10].Nowadays, with the increasing demand of markets, LJF-cultivated locations have spread to Southwest China, including Sichuan and Guangxi provinces [11,12].The processing of LJF in the market primarily includes two methods (oven drying and shade drying).Few studies have compared the effect of the production areas and the processing methods on the fungal community in LJF.
Recent studies reported that herbs could be naturally contaminated by fungi in various procedures, including cultivation, processing, transporting, storage, and marketing [13][14][15].Le et al. detected 153 fungi in medicinal plants collected from Vietnam and identified these microorganisms in seven genera, mainly including Alternaria, Fusarium, and Penicillium [16].In the Kingdom of Saudi Arabia, Al-Hindi et al. analyzed the fungal contamination in 50 herbal samples collected from the local market and indicated that Aspergillus, Penicillium, Fusarium, and Rhizopus were the main contaminated genera [17].In 2016, 187 fungi were isolated by Aiko et al. from 58 out of 63 medicinal herb samples, and 28 fungal strains were found to be toxigenic [18].Moreover, Zheng et al. (2017) detected 126 fungi in 15 different medicinal herbs through morphology and molecular identification.Their results indicated that two species in Penicillium and one species in Eurotium were identified in three LJF samples [19].Therefore, the safety and quality of LJF have increasingly caught public attention in recent years.With the development of DNA sequencing technology, the analysis of fungal community using next-generation sequencing tools has become increasingly acceptable [20,21].DNA metabarcoding has shown potential in monitoring the safety of herbs by accurately detecting the overall fungal composition and diversity [22,23].At present, this technology has been applied in the analysis of the fungal community in herbs, such as Platycladi Semen, Myristicae Semen, and Ziziphi Spinosae Semen, through amplifying the internal transcribed spacer 2 (ITS2) region [24][25][26].
In this study, DNA metabarcoding was applied to investigate the fungal community in LJF samples.On the basis of production area, the samples were divided into five groups.Moreover, the influence of different processing methods on the fungal community in LJF samples was compared.This study is expected to provide a scientific and normative method to for the early warning to supervise fungal and mycotoxin contamination in the LJF industry.

Fungal Diversity in LJF Samples
After chimeric sequences were excluded, a total of 18 LJF samples were detected with 1,301,378 ITS2 sequences, and the average length of the sequences was 322 bp.High-quality sequences were clustered into 504 OTUs. Figure 1A       Six alpha diversity indices were calculated to illustrate the community diversity, richness, evenness, and species coverage in the LJF samples (Table 1).The highest Shannon and lowest Simpson indices were observed in LJFSC1, illustrating that the sample had the highest fungal diversity among the 18 LJF samples.LJFHB3 had the lowest Shannon and highest Simpson metrices, indicating that the diversity in LJFHB3 was lower than in other LJF samples.The ACE and Chao 1 indices of LJFHB3 were lowest, indicating the lowest fungal richness.LJFGX2 had the highest ACE and Chao1 metrices, representing the richest microbial abundance in the 18 samples.The result of Good's coverage in the 18 LJF samples suggested that the fungal community of all samples was sufficiently estimated (>99.9%).In addition, the Shannoneven index of LJFSC1 was higher than those of other samples, illustrating that the highest microbiome evenness was observed in this sample.

Fungal Comparison in LJF Samples from Five Production Areas
Alpha-diversity analysis showed that the LJFHB group had the lowest Shannon index and the highest Simpson index.The ACE and Chao 1 indices of the LJFHB group were lower than those of the other groups.These results illustrated that the fungal richness and diversity in the LJFHB group were lowest among the five groups.The LJFSD group had higher Shannon and Shannoneven and lower Simpson indices, indicating that the fungal diversity was larger than others.The Bar diagram demonstrated that Mucorales in the LJFSD group was lowest among the five groups at the order level.At the family level, the LJFSD group had a higher relative abundance of Pleosporaceae compared to the others.Lefse analysis was performed to compare the differences in fungal community among five LJF groups at various levels, ranging from the phylum level to genus level (Figure S1A,B).At the family level, Dissoconiaceae, Tremellaceae, and Sporocadaceae in the LJFGX group were significantly higher than those in the other groups (p < 0.05).The relative abundances of Naganishia and Trichoderma in the LJFSD group were remarkably more enriched at the genus level (p < 0.05).The NMDS analysis, conducted at the genus level, based on the QIIME calculation illustrated the similarity in the fungal composition in five LJF groups (The lower the abundance of the genus, the darker the orange color; the higher the abundance of the genus, the darker the green color).(B) Data were visualized using Circos.

Fungal Comparison in LJF Samples from Five Production Areas
Alpha-diversity analysis showed that the LJFHB group had the lowest Shannon index and the highest Simpson index.The ACE and Chao 1 indices of the LJFHB group were lower than those of the other groups.These results illustrated that the fungal richness and diversity in the LJFHB group were lowest among the five groups.The LJFSD group had higher Shannon and Shannoneven and lower Simpson indices, indicating that the fungal diversity was larger than others.The Bar diagram demonstrated that Mucorales in the LJFSD group was lowest among the five groups at the order level.At the family level, the LJFSD group had a higher relative abundance of Pleosporaceae compared to the others.Lefse analysis was performed to compare the differences in fungal community among five LJF groups at various levels, ranging from the phylum level to genus level (Figure S1A,B).At the family level, Dissoconiaceae, Tremellaceae, and Sporocadaceae in the LJFGX group were significantly higher than those in the other groups (p < 0.05).The relative abundances of Naganishia and Trichoderma in the LJFSD group were remarkably more enriched at the genus level (p < 0.05).The NMDS analysis, conducted at the genus level, based on the QIIME calculation illustrated the similarity in the fungal composition in five LJF groups (Figure 4A).The stress index indicated that the analysis could be greatly convincing (stress < 0.2).The samples from the LJFHB group and LJFHN groups were close to the samples from the LJFSD group, indicating that these compositions were similar.The LJFGX and LJFSC group could be significantly distinguished with the LJFHB, LJFSD, and LJFHN groups.The result of PcoA analysis, which was conducted at the OTU level, similarly showed that the LJFGX and LJFSC groups varied from the others (Figure 4B).lower the abundance of the genus, the darker the orange color; the higher the abundance of the genus, the darker the green color.)(B) Data were visualized using Circos.

Fungal Comparison in LJF Samples from Five Production Areas
Alpha-diversity analysis showed that the LJFHB group had the lowest Shannon index and the highest Simpson index.The ACE and Chao 1 indices of the LJFHB group were lower than those of the other groups.These results illustrated that the fungal richness and diversity in the LJFHB group were lowest among the five groups.The LJFSD group had higher Shannon and Shannoneven and lower Simpson indices, indicating that the fungal diversity was larger than others.The Bar diagram demonstrated that Mucorales in the LJFSD group was lowest among the five groups at the order level.At the family level, the LJFSD group had a higher relative abundance of Pleosporaceae compared to the others.Lefse analysis was performed to compare the differences in fungal community among five LJF groups at various levels, ranging from the phylum level to genus level (Figure S1A,B).At the family level, Dissoconiaceae, Tremellaceae, and Sporocadaceae in the LJFGX group were significantly higher than those in the other groups (p < 0.05).The relative abundances of Naganishia and Trichoderma in the LJFSD group were remarkably more enriched at the genus level (p < 0.05).The NMDS analysis, conducted at the genus level, based on the QIIME calculation illustrated the similarity in the fungal composition in five LJF groups (Figure 4A).The stress index indicated that the analysis could be greatly convincing (stress < 0.2).The samples from the LJFHB group and LJFHN groups were close to the samples from the LJFSD group, indicating that these compositions were similar.The LJFGX and LJFSC group could be significantly distinguished with the LJFHB, LJFSD, and LJFHN groups.The result of PcoA analysis, which was conducted at the OTU level, similarly showed that the LJFGX and LJFSC groups varied from the others (Figure 4B).

Fungal Comparison in LJF Samples by Using Three Processing Methods
The Shannon and Simpson metrices in LJFYG and LJFHG indicated a higher diversity than that in LJFXY.The fungi were richer in LJFYG and LJFHG than in LJFXY, as revealed by the higher indices of ACE and Chao1.The difference between LJFXY and LJFYG was determined at 99% confidence intervals, as shown in Figure 5A.At the genus level, the relative abundances of Rhizopus and Fusarium in LJFXY were significantly higher than those in the LJFYG.After being processed, the LJF samples had more Aspergillus and Cladosporium, from 4.39% to 4.70% and 1.47% to 2.13%, respectively.In addition, Figure 5B shows the comparison of LJFXY with LJFHG at the genus level.The relative abundances of Rhizopus and Fusarium significantly decreased after being dried in the oven, similar to the processing method wherein samples were dried in the shade.The relative abundance of Aspergillus and Cladosporium in LJFHG was notably lower than that in LJFXY.
those in the LJFYG.After being processed, the LJF samples had more Aspergillus and Cladosporium, from 4.39% to 4.70% and 1.47% to 2.13%, respectively.In addition, Figure 5B shows the comparison of LJFXY with LJFHG at the genus level.The relative abundances of Rhizopus and Fusarium significantly decreased after being dried in the oven, similar to the processing method wherein samples were dried in the shade.The relative abundance of Aspergillus and Cladosporium in LJFHG was notably lower than that in LJFXY.

Fungal Co-Occurrence Analysis in LJF Samples
The interaction of fungi at the genus level was studied via co-occurrence analysis to reveal the microbial diversity of LJF samples (Figure S2).A total of 18 positive and 4 negative correlations were identified among the top 20 detected genera from two phyla, namely Ascomycota and Basidiomycota.The wider the line was, the closer the correlation between the genera was.The correlation between Cladosporium and Alternaria was closer than with Erysiphe, Sporidiobolus, Vishniacozyma, and Filobasidium.Hyphopichia displayed a negative correlation with Alternaria, Cladosporium, Erysiphe, and Sporidiobolus.Fusarium was positively correlated with Clonostachys.Wallemia exhibited a positive interaction with Vishniacozyma and Filobasidium.

Fungal Contaminations in LJF Samples
LJF, as an herbal tea, is easily subjected to various fungal contaminants during planting, harvesting, processing, packaging, transportation, and storage [19].Tea has been consumed as an infusion or a decoction all over the world for thousands of years, and it is filled with soluble and insoluble ingredients.These ingredients may include a number of contaminants, which pose a potential dominant health hazard for humans [27].Fungi, as some of the contaminants in tea, have been studied in recent years.In 2020, Reinholds et al. revealed that 87% of tea samples were contaminated by 32 fungal species, in which five Aspergillus spp.and one Penicillium spp.were predominant [28].Pakshir et al. investigated the fungal contamination in 45 black teas and 15 green teas collected from different brands.The result showed that 89% of black tea samples were contaminated by Aspergillus (66.7%),Penicillium (35.6%),Mocur (20%), and yeast (6.7%), while each green tea sample

Fungal Co-Occurrence Analysis in LJF Samples
The interaction of fungi at the genus level was studied via co-occurrence analysis to reveal the microbial diversity of LJF samples (Figure S2).A total of 18 positive and 4 negative correlations were identified among the top 20 detected genera from two phyla, namely Ascomycota and Basidiomycota.The wider the line was, the closer the correlation between the genera was.The correlation between Cladosporium and Alternaria was closer than with Erysiphe, Sporidiobolus, Vishniacozyma, and Filobasidium.Hyphopichia displayed a negative correlation with Alternaria, Cladosporium, Erysiphe, and Sporidiobolus.Fusarium was positively correlated with Clonostachys.Wallemia exhibited a positive interaction with Vishniacozyma and Filobasidium.

Fungal Contaminations in LJF Samples
LJF, as an herbal tea, is easily subjected to various fungal contaminants during planting, harvesting, processing, packaging, transportation, and storage [19].Tea has been consumed as an infusion or a decoction all over the world for thousands of years, and it is filled with soluble and insoluble ingredients.These ingredients may include a number of contaminants, which pose a potential dominant health hazard for humans [27].Fungi, as some of the contaminants in tea, have been studied in recent years.In 2020, Reinholds et al. revealed that 87% of tea samples were contaminated by 32 fungal species, in which five Aspergillus spp.and one Penicillium spp.were predominant [28].Pakshir et al. investigated the fungal contamination in 45 black teas and 15 green teas collected from different brands.The result showed that 89% of black tea samples were contaminated by Aspergillus (66.7%),Penicillium (35.6%),Mocur (20%), and yeast (6.7%), while each green tea sample showed yeast (66.7%),Aspergillus (60%), Mocur (46.7%),Penicillium (46.7%), and Fusarium (13.3%) [8].Wang et al. observed that Fusarium was the dominant genus in samples collected from the subtropical tea plantations of China [29].Aspergillus, Fusarium, Penicillium, and yeast were the main fungi detected from tea samples in previous studies.In 2020, Liu et al. demonstrated the high abundance of Aspergillus, Penicillium, Xanthomonas, Microcystis, Talaromyces, and Erysiphe in five LJF samples on the basis of ITS sequencing [30].In the present study, the fungal community in 18 LJF samples was investigated through DNA metabarcoding on the basis of ITS2 sequences.Aspergillus, Fusarium, and Penicillium were detected in 18 LJF samples.The relative abundance of Fusarium accounted for 0-11.28%,followed by Aspergillus (0.36-15.62%) and Penicillium (0-1.49%).Much attention has been paid to Fusarium as the primary genus detected in Chinese subtropical tea plantations [31].Aspergillus and Penicillium, which could consequently affect human and animal health, have been reported for years to have toxigenic characteristics.
Ten potential pathogenic fungi belonging to Candida, Malassezia, Kodamaea, Lodderomyces, Schizophyllum, Wallemia, and Mucor were detected via manual BLAST based on 100% ac-curacy.C. tropicalis could cause human diseases under proper conditions, such as bloodstream infections and candidaemia [32,33].C. parapsilosis, K. ohmeri, and L. elongisporus, important nosocomial pathogens, may infect weak patients in hospitals and even threaten life [34][35][36].M. restricta, M. sympodialis, and W. sebi were related to bowel disease, dermatological disorders, and systemic infections [37,38].M. circinelloides could contribute to thrombosis and fatal mucormycosis [39].S. commune has the ability to cause serious infection, such as sinusitis, in patients who are ill [40].As a result, these pathogenic fungi may potentially threaten the safety and quality of LJF products and public health.This work could play a role as an early warning to supervise the fungal community in LJF samples to guarantee human health.

Effect of Processing Methods in LJF Samples
Tea, an everyday drink for some people, can be classified into several categories, including black, white, and green tea, on the basis of different processing methods.During the processing procedure, the fungal community exhibited some remarkable differences in various tea categories [41].Fu brick tea, a post-fermented tea, could be greatly influenced by the microbial change during the manufacturing process procedure, as reported by Li et al. in 2017.The research showed that Aspergillus was the dominant genus among the whole detected genera during manufacturing.The abundance of Aspergillus significantly increased during fermentation and even accounted for 99.99% at the end of the fermentation stage of Fu brick tea [42].In 2020, the fungal community of Cassiae Semen, a roasted tea, was investigated on the basis of the processing methods by Guo et al., involving raw and roasted materials [43].The result indicated that the Penicillium in roasted samples was much more distributed than the raw materials were, and the relative abundances of Aspergillus, Cladosporium, Alternaria, and Rhizopus were significantly higher in raw tea.In 2021, Tong et al. detected the microbial populations in green tea and black tea samples, and in the leaves of the Camellia sinensis, in accordance with different processing methods, including fresh, dry, and withering samples [44].Alternaria, Cladosporium, Aspergillus, and Candida were less abundant in dry and withering samples than in fresh samples.After the samples were dried and withered, the relative abundance of Debaryomyces significantly increased.LJF, as one of the simplest manufacturing teas, was considered as the least contaminated tea group during the procedure [41].Previous studies illustrated that the desiccation stage could significantly decrease the abundance of Aspergillus, Cladosporium, Alternaria, and Rhizopus in the tea samples.In the present study, various degrees of fungi contaminated each LJF sample.LJFHG that was dried in the oven had a lower abundance of Rhizopus, Aspergillus, and Cladosporium than LJFXY (fresh LJF sample).Aspergillus and Cladosporium were more abundant in LJFYG (shade-dried) than in LJFHG.The LJF samples that were dried in the oven had less pathogenic fungi than those that underwent other processing methods.Therefore, LJF should be timely dried in the oven to control the fungal contamination during manufacturing.Moreover, our study indicated that production area also had an impact on the fungal community in LJF samples.Samples from Guangxi and Sichuan, which are at similar latitudes, showed little difference in their fungal composition.The fungal composition in samples from Guangxi and Sichuan was significantly different from those in samples from Hebei, Shandong, and Henan.The above similarities or differences may be related to the climatic conditions of the producing areas.

Sample Collection
A total of 18 LJF samples were collected from different production areas in China.The 15 dried samples from five provinces (Shandong, Hebei, Henan, Guangxi, and Sichuan) were divided into five groups on the basis of production area, namely, LJFSD, LJFHB, LJFHN, LJFGX, and LJFSC.The other samples from Beijing were classified into three groups on the basis of processing method, namely, LJFXY, LJFHG, and LJFYG.LJFXY was not processed.The materials of LJFHG were dried in the oven for 10 h at a temperature of vividly shows that 42 shared OTUs were detected based on production area, with 134 OTUs in LJFSD, 153 OTUs in LJFSC, 260 OTUs in LJFGX, 84 OTUs in LJFHB, and 162 OTUs in LJFHN.Of the OTUs, 25 were unique for LJFSD group, 44 were unique for LJFSC group, 135 were unique for LJFGX group, 13 were unique for LJFHB group, and 26 were unique for LJFHN group.A total of 25 shared OTUs were tested in accordance with the processing methods, with 51 OTUs in LJFYG, 58 OTUs in LJFHG, and 36 OTUs in LJFXY (Figure 1B).

11 Figure 1 .
Figure 1.Venn analysis.(A) The analysis based on the production areas.(B) The analysis based on the processing methods.

Figure 1 .
Figure 1.Venn analysis.(A) The analysis based on the production areas.(B) The analysis based on the processing methods.

Figure 2 .
Figure 2. Percentage of community abundance at the phylum (A), class (B), order (C), and family (D) levels in LJF samples.

Figure 3 .
Figure 3. Fungal composition in LJF samples.(A) Heatmap of the top 30 abundant genera.(Thelower the abundance of the genus, the darker the orange color; the higher the abundance of the genus, the darker the green color).(B) Data were visualized using Circos.

Figure 4 .
Figure 4. Comparison of fungal community based on the production areas.(A) PcoA analysis conducted at the genus level.(B) NMDS diagram estimated at the OTU level.

Table 1 .
Alpha diversity of the fungal community in LJF samples.