Diversity of Endophytic Bacteria Associated with Different Cultivars of Medicinal Plant Arctium Lappa L. and Their Potential for Host Plant Growth Promoting

Purpose Arctium lappa L. is one of the medicinal and food homologous plants in China, which is rich in nutrients and medicinal ingredients. The use of plant growth promoting (PGP) endophytic bacteria is a useful alternative in agricultural production to reduce the use of chemical fertilizers. The aim of this study was to analysis the diversity of endophytic bacteria in different cultivars of A. lappa L. collected from two different geographical locations in China and evaluate PGP traits of the isolates and their potential PGP ability in greenhouse condition. Endophytic bacterial community was investigated by culture-dependent and culture-independent methods. Isolates were screened and investigated for multiple PGP traits, and representative strains were inoculated host seedlings to evaluate the growth promoting effect. A total of 348 endophytic bacteria were obtained and they distributed into four phyla, 30 genera and 73 different species. In addition, high throughput sequencing revealed more abundant bacterial community, including 17 bacterial phyla, and 207 different known genera. A high proportion of PGP traits were detected, including production of indole acetic acid, siderophore, ammonia and phosphate solubilization. Four representative strains with multiple PGP traits of the most prevalent genera were further selected for host inoculation and growth promoting evaluation, and they signicantly increase seedlings length, root length and fresh weight. origin


Introduction
Arctium lappa L. is not only a medicinal plant, but also a delicious vegetable. Its root is edible and rich in amino acids, polyphenols, polysaccharides, volatile oil, avonoids and other nutrients and chemical components ; Chen et al. 2020). At the same time, roots and seeds of Arctium lappa L. are also traditional Chinese medicine, which have antibacterial, anti-tumor, anti-virus, immune regulation and other pharmacological effects (Chan et al. 2011;Li et al. 2019). Arctium lappa L. is widely planted in China, and 80% of total output is exported every year. It is one of the favorite medicinal and food homologous plants, especially in China, Japan and other Eastern Asian countries. A large amount of nitrogen fertilizer is needed in the process of planting and growth of Arctium lappa L. However, the long-term use of chemical fertilizers and continuous cropping will lead to soil hardening, disease and pest increase, continuous cropping barrier and yield decline (Huang et al. 2016). Thus, the use of environmentally friendly biological fertilizer is urgently needed, and the introduction of bene cial microorganisms to plant growth and disease inhibition is recognized worldwide (Bargaz et  For example, endophytic bacteria from Lavandula dentata signi cantly increased adventitious root formation and the rooting capacity of cuttings (Pereira et al. 2016). Endophytic bacteria from the roots of the medicinal plant Alkanna tinctoria Tausch induce alkannin and shikonin production in host hairy roots (Rat et al. 2021). Some endophytic actinobacteria from roots and nodules of chickpea (Cicer arietinum) were found to promote host growth and nodulation in a low N sand-vermiculite system (Vo et al. 2021). In recent years, the composition of abundant and diverse endophytic bacteria has been revealed by pure culture isolation and modern high-throughput sequencing technology. Generally, Bacillus, Streptomyces, Pantoea, Pseudomonas Microbacterium, Burkholderia, Micrococcus and Stenotrophomonas were detected as the most common bacteria, where Bacillus and Pseudomonas are the predominant genera (Afzal et al. 2019;Papik et al. 2020). However, compared with nearly 300,000 species of plants on earth (Ryan et a. 2008), the known composition and diversity of endophytic bacteria is the 'tip of the iceberg'. Therefore, it is necessary to understand the community composition of endophytic bacteria associated with different medicinal plants and their ecological function and application potential in host growth.
The objective of this study is: (1) to isolate and identify endophytic bacteria from different cultivars of Arctium lappa L.; (2) to analyze the composition and diversity of endophytic bacteria from different cultivars and different geographical sources by combined culture-dependent and high-throughput sequencing methods; (3) to evaluate the plant growth-promoting traits of the isolates, such as indole acetic acid (IAA), siderophores production, nitrogen xation and phosphate solubilization, and the effects on the growth of host plants seedlings of selected strains in greenhouse soil.

Isolation of endophytic bacteria
Plant samples were washed with tap water and surface sterilized according to the previously described procedures (Qin et al. 2009). Speci cally, samples were rstly soaked in sterile 0.01% Tween 20 for 1 min, and then soaked in 5% NaClO (effective chlorine content) for 3 min for roots, 2 min for stems and leaves, followed by treating with sterile 2.5% Na 2 S 2 O 3 for 10 min, and then washed with sterile water for 3 times. After that, samples were soaked in 75% ethanol for 2 min for roots and stems, and 1 min for leaves, and nal washing 3-5 times with sterile water. In order to check the reliability of surface disinfection, the sterile water used in the last step was absorbed and spread onto LB and ISP 2 agar media, and the plates were cultured at 30 ℃ for one week to observe whether there were colonies present. If not, the surface disinfection was complete. Then, sterilized tissues were aseptically grinded into smaller fragments using a commercial blender and then spread onto different isolation media. actinomycetes, supplemented with nystatin (50 mg l -1 ) and nalidixic acid (50 mg l -1 ); M8-M10 media were used to isolate bacteria, supplemented with nystatin (50 mg l -1 ). Plates were incubated at 28ºC for 1-4 weeks, and the colonies were obtained and puri ed, and then maintained on slants at 4ºC.

Identi cation of the culturable endophytic bacteria
The isolated strains were rstly classi ed and dereplicated according to their phenotypic characteristics, including colony color, morphology, and presence of sterile hyphae and pigment, and the culture characteristics. Then, selected representative strains were identi ed by 16S rRNA gene sequencing. The total DNA extraction and 16S rRNA gene ampli cation was conducted according to the previous method (Li et al. 2007). The primers 27F (5'-GAGTTTGATCACTGGCTCAG-3') and 1492R (5'-TACGGCTACCTTGTTACGACTT-3') were used to amplify 16S rRNA gene. Then, 16S rRNA gene sequence were compared in NCBI and EzTaxon-e database (Kim et al. 2012) to obtain the nearest type strains and the similarities. All the 16S rRNA gene sequences were deposited at GenBank of NCBI database under the accession numbers MW186019-MW186145.
Endophytic bacterial diversity by high throughput sequencing (HTS) method The number of endophytic bacteria isolated from roots was the largest, thus we continue to analyze the composition of uncultured endophytic bacteria in roots of A. lappa L. by high-throughput sequencing. Healthy roots samples of NB1-NB4 were rstly surface sterilized as described above. Then, about 0.5 g tissues of each sample were used for total DNA extraction by Shanghai Majorbio Biotechnology Co., Ltd. (Shanghai, China).

Plant growth promotion traits evaluation of culturable endophytic bacteria
The PGP traits of the 140 representative strains, including indole acetic acid (IAA) production, siderophore production, ACC deaminase production, ammonia production, phosphate solubilization, nitrogen xation and hydrolase production were determined according to the standard procedures. Nitrogen xation potential was examined by observing the growth on N-free semi-solid JNFb medium (Döbereiner et al. 1995) and both positive PCR ampli cation of nifH gene using the primer pair PolF/PolR and related PCR conditions (Poly et al. 2001). Indole acetic acid (IAA) production was qualitatively evaluated by cultivating strains in minimal salts medium supplemented with 0.5 mg/ml of tryptophan and detection via Salkowsky's reagent (Gordon and Weber, 1951), and quantitative analysis was performed by using the standard curve of IAA standard. Siderophore, ACC deaminase production and phosphate solubilization was checked by using standard Chrome Azurol-S (CAS) assay (Schwyn and Neilands, 1987), DF minimal medium (Penrose and Glick, 2003), and Ca 3 (PO 4 ) 2 solubilization agar medium (Franco-Correaa et al. 2010), respectively. For ammonia production examination, strains were cultured into 5 ml liquid medium containing 10 g/L peptone, incubating at 28°C (120 rpm) for 72 h. Then, 0.5 ml Nessler's reagent (Rashid et al. 2012) was added to check if yellow brown precipitate was produced. The amylase, chitinase, xylanase, protease and cellulase activities was screened by inoculated the strains on the starch agar (Hols et al. 1994), minimal medium agar supplemented with 0.6% (w/v) colloidal chitin and xylan (Chernin et al. 1995 seeds were sterilized with 5% NaClO for 5 min, following washing with sterile water for 3-4 times, and then soaked in endophytic bacterial suspension for 20 h, and the control was soaked in sterile water for 20 h. The treated seeds were directly planted in pots (one seed per pot), which were lled with 200 g sterilized soil. Then, all the pots were transferred to the greenhouse (25 °C/20 °C, 16/8 h light/dark), and each treatment consisted of ten biological replicates.
During seedling growth, the treatment and control groups were watered normally. After 45 days of growth, the growth parameters (total length, fresh weight and root length) of the seedlings were evaluated.

Statistical analysis
The experimental results were expressed by mean values and standard deviation (SD) for three or ve independent replicates as speci ed. Data were analyzed by analysis of variance (ANOVA) and Student's t-tests at P < 0.05. Excel and GraphPad Prism 5.0 were used for data analysis and mapping.

Isolation of endophytic bacteria
A total of 348 endophytic bacteria were isolated from different tissues of Arctium lappa L. using ten different isolation media, of which 108 strains were from NB1 (roots, stems and leaves), 116 strains were from NB2 (roots, stems and leaves), 86 strains were from roots of NB3, and 38 strains were from roots of NB4. There were 230 strains in roots, 68 strains in leaves, and 52 strains in stems, respectively. Comparing the isolation e ciency of different media, we found that M8-M10 media included the most isolated bacteria, with 112, 84, and 65 strains, respectively (Fig. 1A), whereas 87 endophytic bacterial strains were obtained from the other seven different media.
The composition of endophytic bacteria in A. lappa L. was also affected by plant cultivars and geographical origin ( Fig. 2C and D). At the phylum level, there is little difference in bacterial composition, and the most abundant is Actinobacteria, followed by Proteobacteria and Firmicutes (Fig. 2C). At the genus level, 18, 14, 11 and 18 different bacterial genera were obtained from the four samples, respectively ( Fig. 1B and Table S2). However, plant cultivars and geographical factors affect the distribution of culturable endophytic bacteria. The genera Bacillus, Microbacterium, Pseudomonas and Streptomyces were common in the four samples. Among the three A. lappa L. cultivars sampled from Xuzhou, all the most dominant genera were Microbacterium and Streptomyces. However, abundant Promicromonospora was isolated from NB3 cultivar and rarely obtained from other samples. The genera Arthrobacter, Microbacterium and Pseudoclavibacter were found abundant in the Harbin sample NB4, which is different from the mainly composition of three Xuzhou samples (Fig. 2D).
Diversity of unculturable endophytic bacteria in roots of A. lappa L.
A total of 145,525 quality-ltered reads were obtained from four root samples (27,573 per sample). Root of NB3 harbors the most abundant OTUs (494) at the species level. The number of OTUs in other samples were 332, 449, and 221 for NB1_R, NB2_R and NB4_R, respectively. Diversity coverage estimations indicated more than 99.0% of bacteria was sequenced. Moreover, rarefaction analyses indicated that the curves nearly reached asymptotes, suggesting that measured diversity was nearly saturated (Fig. 3A). Alpha diversity indices were calculated and indicated that community diversity differentiate across the for roots samples, with NB2 and NB3 roots harbored more abundant bacterial endophytes (Table 1). A total of 100 OTUs were common present across four samples. In contrast, OTUs that were exclusive to each root tissue ranged from 43 to 141. NB3_R harbored the highest numbers of unique OTUs, followed by NB2_R and NB4_R (Fig. 3B).
The distribution of endophytic bacteria at the phylum level in different root samples is shown in Fig. 4C. A total of 17 phyla of bacteria were detected, among which Proteobacteria (47.98%-96.72%) was the most dominant, followed by Firmicutes, Actinobacteria, Bacteroidetes and Cyanobacteria. The phyla Firmicutes (23.03% and 32.02%) and Actinobacteria (6.843% and 12.01%) in NB2_R and NB3_R samples showed much higher relative abundance than other two samples (Fig. 3C). The composition of endophytic bacteria in different roots varied greatly at genus level. A total of 207 different known genera were detected by HTS technique. In addition to the unclassi ed genera, the genus Pantoea accounted for 53.19% in NB1, and Pseudomonas was the most abundant genus in NB2 and NB4 samples, accounting for 19.24% and 76.94% respectively, whereas Hyphomicrobium was more abundant in NB2, accounting for 7.35% (Fig. 3D). Except for the unclassi ed genera, the top ten core genera in each root tissue are shown in Table 2. We found that Pseudomonas, Bacillus, Pandora and Streptomyces were distributed in almost all samples, and they were the most common genera in A. lappa L. roots.
Phylogenetic a liation of the 16S rRNA gene sequences revealed more abundant endophytic bacteria than those obtained by pure culture method.
Speci cally, only 4 of the 17 phyla detected by HTS technique were obtained by cultivation. At the genus level, 83-154 known genera were found in four kinds of roots, far more than the maximum of 18 genera obtained by pure culture technique (Fig. 1B). Many genera detected by HTS were also obtained by pure culture, including Amycolatopsis, Brachybacterium, Enterobacter, Methylobacterium, Microbacterium, Nonomuraea, Promicronospora, Rhizobium, etc. The core genera, including Pseudomonas, Pantoea, Arthrobacter, Bacillus, Streptomyces, Pseudomonas and Rhizobium, were also cultured by culture-dependent methods. However, some genera were not detected by high-throughput sequencing, such as Achromobacter, Cedecea, Cellulosimicrobium, Nocardia, Plantibacter, Pseudoclavibacter, Rhodococcus, Sphingobacterium and Staphylococcus, which may be related to the quality of DNA extraction and the bias of PCR ampli cation in high-throughput sequencing.

Plant growth promotion traits of culturable endophytic bacteria
After checking the PGP traits of the 140 strains, the positive for the ACC deaminase and ammonia production was up to 55% (77 strains) and 54.3% (76 strains), respectively, followed by siderophore production (48.6%, 68 strains), nitrogen xation (46.4%, 65 strains), IAA production (41.4%, 58 strains), and phosphate solubilization (26.4%, 37 strains). Comparing the origin of these strains, we found that the endophytic bacteria from NB1 showed the most PGP characteristics, almost all more than 20 strains, followed those from NB2 and NB3, and the number of active strains from NB4 was the least (Fig. 4A). The positive results of representative strains are shown in Fig. 4C. For IAA quantitative detection, the yield of strain Pantoea SP. WY068 was the highest, reaching 37.8 mg/L. In addition, 47 isolates (33.6%) possessed amylase activity, 46 isolates (32.9%) possessed cellulase activity, and 45 isolates (32.1%) possessed protease activity. Whereas xylanase and chitinase producers were 28 isolates (20%) and 22 strains (15.7%), respectively. In comparison, NB3 possessed the largest number of enzymes producing bacteria, more than 10 strains for each hydrolytic enzyme (Fig. 4B). The enzyme activity screening of representative strains is shown in Fig. 4D.

Plant growth promoting activity of selected bacterial endophytes
Four representative endophytic bacteria, WY035 (Bacillus sp.), WY068 (Pantoea sp.), WY358 (Microbacterium sp.), and WY297 (Pseudomonas sp.), which showed multiple PGP characteristics (Table 3) of the most prevalent genera were selected for PGP evaluation. After seeds inoculation and 45 days growth in soil, seedlings growth was evaluated. There was no signi cant difference in seed germination rate after the treatment of the four strains. However, the results showed that all the strains stimulated the growth of A. lappa L. seedlings. Endophytes treated seedlings showed signi cant increase (P < 0.05) in seedlings length, root length and fresh weight (Fig. 5). The former three strains exhibited the highest PGP activities, and their results in pots were shown in Fig. 5D.
Seedlings length was increased by 68.3% and 69.4%, respectively, after WY035 and WY068 inoculation (Fig. 5A). In addition, isolates WY068 and WY035 also greatly increased the root lengths by 133.9% and 94.3%, respectively (Fig. 5B). Furthermore, the fresh weight of the seedlings was increased at least by 4.5folds in the presence of the endophytes (Fig. 5C).

Discussion
Plants harbor abundant endophytes, and a variety of endophytic bacteria have been detected by culture-dependent and culture-independent methods. In this study, we isolated more than 300 endophytic bacteria strains, including 30 genera and a potential new species, indicating that A. lappa L. harbors abundant and diverse endophytic bacteria. According to the literature, there are few reports of endophytic bacterial diversity from a medicinal plant with such high abundance. Taxonomic compositional analysis indicated that these strains distributed into bacterial phyla Actinobacteria, Proteobacteria, Firmicutes and Bacteroidetes. The composition of the endophytes at the phylum level is consistent with various current results (Hardoim et  , indicating that they are common and widely distributed in different plants. Some rare genera, such as Leucobacter, Cedecea, Pseudoclavibacter and Plantibacter have rarely been reported as pure cultures of endophytic bacteria, which may be related to host selection and recruitment effect of endophytic bacteria (Mina et al. 2020). High throughput sequencing technology provides convenience for us to understand the composition of endophytic bacteria. As many as 207 known endophytic bacterial genera some unclassi ed taxa were detected by HTS technique in this study, which indicated that the diversity of endophytic bacteria in roots was far more than that of pure culture. Only 10.1% (21 genera) of the 207 genera were also obtained by culture-dependent method.
The dominant genera found by cultivation are not completely consistent with HTS technique. For example, HTS revealed abundant OTUs of Pantoea and Pseudomonas, but the number of pure cultures was less. This is related to the selection of isolation method, media used, primer preference for PCR and the extraction e ciency of total environmental DNA. This disparity is consistent with some recent results (Qin et  In this study, cultivars and geographical factors greatly affected the composition of endophytic bacteria in A. lappa L., and the difference was more obvious at the genus level. Even for the three cultivars grown in the same eld, the dominant genera were different (Fig. 4). This result is consistent with the recent discovery that the composition of endophytic bacteria in different cultivars of mulberry is different (Xu et al. 2019). Besides, 66.1% of the strains were isolated from the root tissues of A. lappa L., which was similar to many other reports (Qin et al. 2009;Tian and Zhang, 2017). Studies have con rmed that root endophytes mainly come from the rhizosphere soil, and then become endophytes through the recruitment of plant root exudates (Martínez-Romero et al. 2020). Because roots of A. lappa L. is the edible part, elucidating the diversity of endophytic bacteria in the roots is helpful to understand their functions and potential roles in plant microecosystem.
Studies have con rmed that endophytic bacteria can promote plant growth through a series of direct and indirect mechanisms, such as nitrogen xation, IAA secretion, dissolution of phosphorus and potassium, production of siderophore and various hydrolases (Qin et al. 2011;Rana et al. 2020). We found that the isolated endophytic bacteria showed a variety of PGP traits, especially the numbers of IAA and siderophore producers, and nitrogen xation endophytes were more than 40% of the detected strains (Fig. 4). The results showed that the endophytic environment of A. lappa L. is rich in endophytic bacteria with plant growth promoting function, and the relationship between endophytes and host is worth exploring. We further selected four representative endophytic bacteria of the dominant genera (Bacillus, Pantoea, Pseudomonas and Microbacterium) for seed inoculation, and found that they can signi cantly promote host growth (Fig. 5), which is mainly due to their multiple PGP characteristics. It is worth noting that the community of endophytic bacteria with multiple PGP characteristics are very abundant in this study, which indicates that it is indeed the microbiota, rather than a single strain, may play a more important role in host growth promoting. At present, the use of simpli ed and synthetic bacterial community to study plant-microbiome interaction is still in its infancy (Vorholt et al. 2017). However, synthetic microbiome is a good strategy. There They found that the simpli ed microbiome could promote the growth of maize seedlings and signi cantly inhibit the colonization of pathogen Fusarium verticillioides. Simpli ed microbiota construction is an important direction of plant microbiome research in the future. Therefore, a simpli ed microbiome of Arctium lappa L. can be constructed for host growth promotion evaluation and eld inoculation veri cation in the further study.

Conclusion
In this study, the composition of endophytic bacteria in different cultivars of Arctium lappa L collected from two different geographical locations was analyzed in detail by culture-dependent and 16S rRNA gene high-throughput sequencing techniques. Overall, 348 endophytic bacteria were obtained, and they belonged to four bacterial phyla (Actinobacteria, Proteobacteria, Firmicutes and Bacteroidetes). Results demonstrated that endophytic bacteria associated with A. lappa L., especially in the roots were abundant. In addition, a high proportion of bene cial bacteria with potential growth promoting function were found in the endophytic bacteria, and four representative strains promoted host seedling growth signi cantly in the pot experiments. Our results provided good resources for the subsequent development of biological fertilizers for A. lappa L. growth, but the actual promotion effect in the eld still needs to be evaluated. Besides, the detailed growth promoting mechanism of single strain and the construction of synthetic bacterial community need to be further studied.