E ﬀ ects of the Biofertilizer OYK ( Bacillus sp.) Inoculation on Endophytic Microbial Community in Sweet Potato

: Sweet potato ( Ipomoea batatas L.) grows well even in infertile and nitrogen-limited ﬁelds, and endophytic bacterial communities have been proposed to be responsible for this ability. Plant-growth-promoting bacteria are considered eco-friendly and are used in agriculture, but their application can interact with endophytic communities in many ways. In this study, a commercial biofertilizer, OYK, consisting of a Bacillus sp., was applied to two cultivars of sweet potato, and the e ﬀ ects on indigenous endophytic bacterial communities in ﬁeld conditions were examined. A total of 101 bacteria belonging to 25 genera in 9 classes were isolated. Although the inoculated OYK was not detected and signiﬁcant plant-growth-promoting e ﬀ ects were not observed, the inoculation changed the endophytic bacterial composition, and the changes di ﬀ ered between the cultivars, as follows: Novosphingobium in α -Proteobacteria was dominant; it remained dominant in Beniharuka after the inoculation of OYK, while it disappeared in Beniazuma, with an increase in Sphingomonas and Sphingobium in α -Proteobacteria as well as Chryseobacterium and Acinetobacter in Flavobacteria. The behavior of Bacilli and Actinobacteria also di ﬀ ered between the cultivars. The Shannon diversity index ( H ) increased after inoculation in all conditions, and the values were similar between the cultivars. Competition of the inoculant with indigenous rhizobacteria and endophytes may determine the fates of the inoculant and the endophytic community.


Introduction
Modern agriculture systems are being intensified through the use of various technologies to achieve maximum efficiency and high qualify products to meet the growing global demand for food supply [1]. At present, as a part of agricultural intensification, crop production depends on the large-scale use of chemical fertilizers [2]. However, the intensive use of chemical fertilizers can result in considerable negative environmental impacts and pollution [3]. Therefore, an alternative strategy is urgently needed to establish sustainable agriculture and ecological balance in agro-ecosystems.
Plant-growth-promoting rhizobacteria (PGPR) are free-living soil bacteria that enhance plant growth by colonizing the rhizosphere [4]. PGPR regulate nutritional and hormonal balance, produce chemical fertilizer (N:P 2 O 5 :K 2 O = 4:8:15 g/m 2 ) was applied before planting. The plants were cultivated from June to November in 2015 with drip irrigation (Super Typhoon NETAFIM Co., Tel Aviv, Israel).

Sample Collection and Isolation of Endophytic Bacteria
At harvest, the fresh weights of the shoots and tubers of each sweet potato plant were measured. Culturable endophytes of sweet potato tubers were examined; among the plant parts, the highest population was observed in tubers in our previous study [34]. The surface of each tuber sample was washed with running tap water for 10 min and cut longitudinally with a sterilized knife at its middle part after wiping off the water with a paper towel. Then, the cut surface was stamped on modified MR agar medium, with and without the supplementation of ammonium nitrate as a nitrogen source [35] in a petri dish. The ingredients of the media are listed in Supplementary materials Table S1. The efficiency of the washing procedure was evaluated by stamping the surface of the washed tubers on agar media. After incubation for 2 days at ca. 26 • C, all the bacterial colonies were transferred to either N-supplemented or N-free MR media for purification and then grouped based on their morphologies on the two media. Based on their relative abundance, 1-3 representative isolates from each group, comprising 30-81% of total isolates, were selected for further analysis (Table 1).

Genetic Analysis of Endophytes
Genomic DNA was extracted from each isolate, as described by Saeki et al. [36], with slight modifications, and used as a template for PCR for the amplification of the partial 16S rRNA gene sequence. As an indication of the dinitrogen-fixing potential of the isolates, nifH genes, which encode nitrogenase reductase, were PCR-amplified, for which a small amount of culture was directly used as a template. The primers used were fD1 and rP2 [37] and PolF and PolR [38] for the 16S rRNA and nifH genes, respectively. The components of the PCR master mixtures and the PCR running conditions are summarized in Supplementary materials Table S2. PCR products were purified and subjected to PCR cycle sequencing, according to the procedures described previously [39].
The closest sequence in the database (https://www.ddbj.nig.ac.jp/) was determined by a BLAST [40] search, and multiple sequence alignments were constructed using ClustalW 2.1 [41]. Alignments were manually edited, and phylogenetic trees with the related reference genes were constructed using ClustalW 2.1 with the neighbor-joining method.

Analysis of the Community Structure of Endophytes
Based on the results of the BLAST search and phylogenetic analysis, relative abundance (%) was calculated according to the class and genus of the identified bacteria for each sample, reflecting the relative abundance on the plate (Table 1). These results were used to analyze the effects of OYK inoculation, the difference between the presence and absence of a nitrogen source in the medium, and the two sweet potato cultivars on the community structure of the endophytes. Principal component analysis (PCA) was applied on a genera basis using IBM SPSS Statistics ver. 25 (IBM Co., Armonk, NY, USA).

Nucleotide Sequence Accession Numbers
The sequence data generated in this study were deposited in the DDBJ Nucleotide Submission System under the accession numbers LC583148 to LC583248.

Statistical Analysis
Statistical analysis of the sweet potato cultivation data was performed using Student's t-test. The Shannon diversity index (H') was calculated based on the identified genus to characterize the diversities in the endophytic bacterial communities.

Effects of OYK Inoculation
In terms of the dry weights of shoots and tubers, the growth of sweet potato cultivar Beniharuka was better than that of Beniazuma, and there was no significant difference between samples with and without OYK inoculation in either cultivar ( Figure 1). Based on the results of the BLAST search and phylogenetic analysis, relative abundance (%) was calculated according to the class and genus of the identified bacteria for each sample, reflecting the relative abundance on the plate (Table 1). These results were used to analyze the effects of OYK inoculation, the difference between the presence and absence of a nitrogen source in the medium, and the two sweet potato cultivars on the community structure of the endophytes. Principal component analysis (PCA) was applied on a genera basis using IBM SPSS Statistics ver. 25 (IBM Co., Armonk, NY, USA).

Nucleotide Sequence Accession Numbers
The sequence data generated in this study were deposited in the DDBJ Nucleotide Submission System under the accession numbers LC583148 to LC583248.

Statistical Analysis
Statistical analysis of the sweet potato cultivation data was performed using Student's t-test. The Shannon diversity index (H') was calculated based on the identified genus to characterize the diversities in the endophytic bacterial communities.

Effects of OYK Inoculation
In terms of the dry weights of shoots and tubers, the growth of sweet potato cultivar Beniharuka was better than that of Beniazuma, and there was no significant difference between samples with and without OYK inoculation in either cultivar ( Figure 1).

Isolation of Endophytic Bacterial Strains
Originally, 269 bacterial colonies appeared on the agar plates in total, of which 232 strains were successfully isolated. On the basis of their observed morphologies on the modified MR agar medium, with and without nitrogen supplementation, the isolates were grouped into 6-17 groups in each sample. Based on their relative abundance, 1-3 representative isolates were selected from each group, comprising 30-81% of the original isolates; as a result, 109 isolates were selected, in total, for further analysis (Table 1).

Genetic Analysis of Endophytes
Among the 109 selected endophytic bacterial isolates, 101 strains were successfully sequenced for the partial 16S rRNA gene. The results of the closest relatives in the DDBJ database are presented in Supplementary materials Table S3 and Figure S1 and summarized in Table 2 and Figure 2. The

Isolation of Endophytic Bacterial Strains
Originally, 269 bacterial colonies appeared on the agar plates in total, of which 232 strains were successfully isolated. On the basis of their observed morphologies on the modified MR agar medium, with and without nitrogen supplementation, the isolates were grouped into 6-17 groups in each sample. Based on their relative abundance, 1-3 representative isolates were selected from each group, comprising 30-81% of the original isolates; as a result, 109 isolates were selected, in total, for further analysis (Table 1).

Genetic Analysis of Endophytes
Among the 109 selected endophytic bacterial isolates, 101 strains were successfully sequenced for the partial 16S rRNA gene. The results of the closest relatives in the DDBJ database are presented in Supplementary materials Table S3 and Figure S1 and summarized in Table 2 and Figure 2. The isolates belonged to 25 bacterial genera in 9 classes, which showed 97-100% homology. Among the 101 identified bacterial strains, 55 representative strains from each genus in each sample were subjected to PCR for the nifH gene; however, none of the strains produced positive amplification, with Bradyrhizobium elkanii USDA 94 used as a positive control.
Horticulturae 2020, 6, x FOR PEER REVIEW 5 of 12 isolates belonged to 25 bacterial genera in 9 classes, which showed 97-100% homology. Among the 101 identified bacterial strains, 55 representative strains from each genus in each sample were subjected to PCR for the nifH gene; however, none of the strains produced positive amplification, with Bradyrhizobium elkanii USDA 94 used as a positive control. Beniharuka (H), inoculated with OYK as PGPR, compared with the control. Bacteria were cultured using a modified MR medium, with and without a supplemental nitrogen source. Table 2. Relative abundance (%) of endophytes from two cultivars of sweet potato, with and without OYK inoculation as PGPR. Bacteria were cultured using a modified MR medium, with and without a supplemental nitrogen source. 21  60  29  55  62  69  50  36  Novosphingobium  -50  -45  38  54  25  36  Rhizobium  -10  6 -
To further elucidate the influence of the OYK inoculation, PCA was conducted to evaluate the relative abundance of the endophytic genera in Table 2. The first and second component factors explained 61.1% and 13.8% of the variation, respectively (Figure 3). All control samples, including both cultivars and both media conditions, were positioned close to each other, while the OYK-inoculated samples were positioned farther apart for each cultivar, especially in Beniazuma. The effects of the presence or absence of nitrogen in the media were not apparent.  To further elucidate the influence of the OYK inoculation, PCA was conducted to evaluate the relative abundance of the endophytic genera in Table 2. The first and second component factors explained 61.1% and 13.8% of the variation, respectively (Figure 3). All control samples, including both cultivars and both media conditions, were positioned close to each other, while the OYKinoculated samples were positioned farther apart for each cultivar, especially in Beniazuma. The effects of the presence or absence of nitrogen in the media were not apparent.   Table 2.

Diversity of Endophytes
Shannon diversity indices (H), calculated on the genus level, were increased with the inoculation of OYK in all conditions (Figure 4 and Supplementary materials Figure S2). The increase appeared to be larger in endophytic communities that were isolated using nitrogen-free media, although the indices were similar among the control samples. No difference between the cultivars was apparent.
Horticulturae 2020, 6, x FOR PEER REVIEW 7 of 12 control (C). Bacteria were cultured using a modified MR medium, with (+) and without (−) a supplemental nitrogen source. PCA was performed based on the bacterial genera in Table 2.

Diversity of Endophytes
Shannon diversity indices (H), calculated on the genus level, were increased with the inoculation of OYK in all conditions (Figure 4 and Supplementary materials Figure S2). The increase appeared to be larger in endophytic communities that were isolated using nitrogen-free media, although the indices were similar among the control samples. No difference between the cultivars was apparent.
The endophytic community structure has been reported to be determined by several factors, such as plant genotype, soil type [55], and environmental conditions, as well as stochastic sampling
The endophytic community structure has been reported to be determined by several factors, such as plant genotype, soil type [55], and environmental conditions, as well as stochastic sampling factors [56]. In the present study, analysis of the bacterial endophytes of sweet potato revealed that Proteobacteria was the dominant phylum in the communities, followed by Flavobacteria, Sphingobacteria, Actinobacteria, and Bacilli. α-Proteobacteria was the dominant class in Proteobacteria, followed by βand γ-Proteobacteria (Table 2). In previous studies of sweet potato endophytes, Proteobacteria, including α-, β-, and γ-Proteobacteria, Flavobacteria, Actinobacteria, and Bacilli were also predominant among isolates [27,28,57]. These results suggest that the endophytic community of sweet potato consists of bacteria belonging to common phyla.
Almost all of the detected genera in Proteobacteria, Actinobacteria, and Bacilli have been reported as endophytes in sweet potato [27,28,57] except for Novosphingobium sp., which was the dominant genus in most samples. The other dominant genera in our study, Chryseobacterium sp., Acinetobacter sp., Mucilaginibacter sp., and Sphingobacterium sp., have not been reported as endophytes. The genera in Flavobacteria and Sphingobacteria were isolated from the cultivar Beniazuma, suggesting that these isolates were sweet potato cultivar-dependent. Differences in endophytic and rhizosphere bacterial communities among sweet potato cultivars have also been demonstrated [27,58]. On the other hand, the common dominant genera in the other studies, Enterobacter sp., Pantoea sp., Luteibacter sp., Herbaspirillum sp., and Curtobacterium sp., were not isolated in our study, suggesting the presence of diverse bacterial endophytes of sweet potato, with some common genera.
The inoculation of OYK changed the composition of the indigenous bacterial endophytic communities on both the phylum and genus levels, though OYK itself failed to maintain a population as an endophyte. The effects were similar between N(+) and N(−) media, while they were different between the Beniazuma (A) and Beniharuka (H) cultivars, especially for Novosphingobium sp., which was dominant in all control samples and disappeared in Beniazuma (A) while remaining predominant in Beniharuka (H). Flavobacteria and Sphingobavteria in Beniazuma (A) only appeared after the inoculation of OYK, which could have caused the change in the community structures found in PCA ( Figure 3). Although only one sample of the sweet tuber was used for each cultivar and media condition, the closer positions of the control samples indicate that variability in the community structures of the control samples was within a certain range and that the different positions in PCA were caused by the inoculation of OYK. These results suggest that interactive endophytic bacterial behavior might be influenced by the cultivar of sweet potato. It has been reported that the plant cultivar and genotype affect communities of rhizobacteria, presumably as a result of competition for different root exudates [59][60][61]. Differences in a rhizobacterial community might affect the corresponding endophytic community as a result. Germida et al. [62] compared rhizoplane and endophytic bacteria strains that were isolated from canola plants and suggested that endophytes are a subset of the rhizoplane community. Additionally, differences in nutritional compositions of endophytic environments will also affect the community through competition.
In a seed and soil inoculation experiment with Bacillus spp., the Bacillus inocula failed to establish as endophytes in broccoli roots, as in our study, and the main effects of the Bacillus inoculation were a reduction in Lysobacter and Acidovorax and an increase in Acinetobacter, as analyzed by metagenomic sequencing [31]. The authors also reported that the addition of B. amyloliquefaciens influenced the endophytic microbial community: the most common Pseudomonas endophytes decreased in abundance, accompanied by an increase in Dyadobacter, Variovorax, Tahibacter, and Sphingomonas. In contrast, the inoculation of B. cereus and B. subtilis did not affect the population of Pseudomonas though it changed the endophytic community composition of minor genera. Although the genera affected by the Bacillus inoculation were different from those in our study, the results obtained by culture-dependent and -independent studies suggest that a microbial inoculation can change an endophytic microbial community, even if the inoculant cannot establish a population as an endophyte. As many studies have shown the importance of endophytes for plant growth promotion, elucidating the interaction mechanisms is an essential line of research.
Although Bacillus spp. have been reported as indigenous endophytes in sweet potato [28,57,58] and in other crops such as tomato [63], banana [64], canola [62], and switchgrass [65], the inoculated OYK and Bacillus spp. strains [31] could not establish populations as endophytes. On the other hand, the inoculation of endophytic Bacillus subtilis, isolated from wheat, could establish a population in wheat root and showed potential as a biological control against plant pathogens [66]. Changes in the compositions of plant metabolites and root exudates that would be caused by OYK might directly change indigenous rhizospheric and endophytic microbial communities and/or might indirectly prevent the successful colonization of OKY due to competition with microbial communities for compounds. As OYK was isolated from the soil, the endophytic potential of an inoculant, whether it was originally isolated as an endophyte, seems to be important.
The Shannon diversity index (H) of the isolated endophytic community increased with OYK inoculation (Figure 3). The tendency was the same as that in the results obtained by Gadhave et al. [31], who also reported an increase in diversity in both Bacillus amyloliquefaciensand mixed Bacillus spp.-treated sprouting broccoli, examined by a culture-independent metagenomic approach. In both studies, using different approaches, the number of genera identified increased with the inoculation; however, the mechanisms are still unclear.
Supplementary Materials: The following are available online at http://www.mdpi.com/2311-7524/6/4/81/s1, Figure S1: Phylogenetic tree of endophytes of sweet potato cultivars, Beniazuma (A) and Beniharuka (H), inoculated with PGPR, OYK compared with control, using modified MR medium supplemented with and without nitrogen source based on partial 16S rRNA gene sequences, Figure S2: Relative genus composition of endophytes of sweet potato cultivars, Beniazuma (A) and Beniharuka (H), inoculated with PGPR, OYK compared with control, using modified MR medium supplemented with and without nitrogen source, Table S1: Ingredients of modified MR (N-free MR) agar medium, Table S2: PCR ingredients for amplification of 16S rRNA and nifH genes, Table S3: Closest relatives of endophytic bacterial strains from two cultivars of sweet potato inoculated with and without PGPR, OYK, using modified MR medium supplemented with and without nitrogen source.