Effect of Co-Inoculation of Bacillus sp. Strain with Bacterial Endophytes on Plant Growth and Colonization in Tomato Plant ( Solanum lycopersicum )

: Colonization of a biofertilizer Bacillus sp. OYK strain, which was isolated from a soil, was compared with three rhizospheric and endophytic Bacillus sp. strains to evaluate the colonization potential of the Bacillus sp. strains with a different origin. Surface-sterilized seeds of tomato ( Solanum lycopersicum L. cv. Chika) were sown in the sterilized vermiculite, and four Bacillus sp. strains were each inoculated onto the seed zone. After cultivation in a phytotron, plant growth parameters and populations of the inoculants in the root, shoot, and rhizosphere were determined. In addition, effects of co-inoculation and time interval inoculation of Bacillus sp. F-33 with the other endophytes were examined. All Bacillus sp. strains promoted plant growth except for Bacillus sp. RF-37, and populations of the rhizospheric and endophytic Bacillus sp. strains were 1.4–2.8 orders higher in the tomato plant than that of Bacillus sp. OYK. The plant growth promotion by Bacillus sp. F-33 was reduced by co-inoculation with the other endophytic strains: Klebsiella sp. Sal 1, Enterobacter sp. Sal 3, and Herbaspirillum sp. Sal 6., though the population of Bacillus sp. F-33 maintained or slightly decreased. When Klebsiella sp. Sal 1 was inoculated after Bacillus sp. F-33, the plant growth-promoting effects by Bacillus sp. F-33 were reduced without a reduction of its population, while when Bacillus sp. F-33 was inoculated after Klebsiella sp. Sal 1, the effects were increased in spite of the reduction of its population. Klebsiella sp. Sal 1 colonized dominantly under both conditions. The higher population of rhizospheric and endophytic Bacillus sp. in the plant suggests the importance of the origin of the strains for their colonization. The plant growth promotion and colonization potentials were independently affected by the co-existing microorganisms.


Introduction
Plant growth-promoting rhizobacteria (PGPR) are becoming more widely accepted in intensive agriculture to enhance sustainable agricultural production in various parts of the world [1]. PGPR contain a diverse range of bacteria and several mechanisms have been proposed though they are not fully understood [2]. In sustainable agricultural practices using PGPR, inoculation techniques for their colonization at the rhizosphere is critical [3]; therefore, a further understanding of the interactions of PGPR with plant and indigenous rhizobacteria is essential.
Bacillus spp. have been recognized as one of the most important PGPR and widely used for sustainable agriculture as biofertilizers and/or antagonists against plant diseases [4][5][6][7][8].
Bacillus spp. have also received considerable attention because of their benefits over other PGPR in producing stable formulations [6,9] and stability in rhizosphere soil in semi-arid deserts [10]. In addition, Bacillus spp. exhibit a significant reduction in disease incidence on various crops by inducing systemic resistance [11,12] and by forming biofilm on root surfaces [13].
In our previous study, when the commercial biofertilizer OYK consisting of the Bacillus sp. strain was applied to sweet potato, no significant plant growth-promoting effect was observed, and the inoculated Bacillus sp. strain was not detected in the plant tubers. The possible reasons were due to competition of the inoculant against indigenous rhizobacteria and endophytes, and a lack of endophytic potential of the inoculant, which was originally isolated from soil [14]. As many endophytic Bacillus strains have been reported in several plants [15][16][17][18][19][20][21][22], it is assumed that endophytic bacteria have some colonization strategies in interaction with plants.
In addition to the individual colonizing ability of PGPR, interactions with other co-existing bacteria would be important to determine the colonization and plant growthpromoting potential. Synergetic effects of the inoculation with the other PGPR have been reported in maize [23], cotton [24], ryegrass [25], strawberry [26], and cucumber [27]. On the other hand, negative interactions with co-existing bacteria should also be considered. They inhibited the colonization of inoculants in sugarcane [28] and reduced the plant growth-promoting effects in tomato plant [29,30].
For efficient and practical use of PGPR, it is essential to understand its colonizing behavior and abilities to compete with co-existing bacteria. Though several studies have been reported on the effects of co-inoculation with multiple bacteria on plant growth, their effects on colonization have not been extensively studied yet. The aim of this study was to evaluate the colonization properties of Bacillus sp. OYK, which was isolated from a soil, in relation to its origin by comparing it with those of the other Bacillus sp. strains isolated from plant endosphere and rhizosphere, and then to elucidate the effects of co-inoculation of the endophytic Bacillus sp. strain with the other endophytes on their colonization and plant growth-promoting activities.

Bacterial Strains
In addition to Bacillus sp. OYK, three strains of Bacillus sp.: two strains (Bacillus sp. RF-12 and RF-37) isolated from the rhizosphere of sweet potato and another one (Bacillus sp. F-33) as an endophyte of the same plant cultivated in Japan [16], and three strains of endophytes: Herbaspirillum sp. Sal 6, Klebsiella sp. Sal 1, and Enterobacter sp. Sal 3, isolated from Nepalese sweet potato [15], were used in this study (Table 1). Table 1. Bacterial isolates used in this study [5,6].

Plant Growth Promotion and Colonization of Bacillus sp. Strains in Tomato Plant
To prepare the bacterial inoculum, each Bacillus sp. strain was cultivated in Modified Rennie (MR) [31] liquid medium with shaking at 150 rpm at 26 • C for 3 days. The culture was washed twice with sterilized distilled water by centrifugation at 10000× g at 4 • C for 10 min, and the cell pellet was resuspended with sterilized distilled water at 10 8 colony forming units (CFU)/mL to prepare an inoculum based on OD-CFU/mL correlated linear equations prepared for each strain.
In this study, we used tomato as a test plant due to the difficulty in preparing bacteriafree plants in sweet potato. Tomato seeds (Solanum lycopersicum L. cv. Chika F1 hybrid, Takii & Co., Ltd., Kyoto, Japan) were surface sterilized with 70% ethanol for 1 min followed by 1% sodium hypochlorite with 3-4 drops of Tween-20 for 13 min and washed 7-8 times with sterilized distilled water. The seeds were sown in the sterilized vermiculite in a Leonard jar [32] supplied with the sterilized Hoagland solution [33], and 1 mL of the inoculum was added onto the seed zone. The jar was put in a ventilated (<0.2 mm pore size) transparent plastic bag (Sun bag, Sigma-Aldrich, Tokyo, Japan), and after thinning out to one plant per jar, the tomato plant was aseptically cultivated in a phytotron (Model-LH 220S, Nippon Medical & Chemical Instruments Co., Ltd., Osaka, Japan) at 28/25 • C (16h/8h, day/night) for 24 days. An autoclaved culture was used as a control, and the experiment was conducted twice, using three plants for each treatment.
After cultivation, the tomato crop was harvested, and the fresh weight and length of the root and shoot were measured. Then, the population of the inoculated strains in the root, shoot, and rhizosphere was determined using two plants for each treatment. A rhizosphere sample was prepared by dipping and gently shaking the roots in sterilized distilled water. After washing the plant surface 6-7 times with sterilized distilled water, the root and shoot samples were separated and macerated with sterilized distilled water using a sterilized mortar and pestle, and the samples were subjected to dilution plating for the determination of CFU/g. At the same time, an aliquot of the final washing solution was directly plated, and no colony was observed. The inoculation experiment was conducted twice.

Effect of Co-Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 with the Other Endophytic Strains in Tomato Plant
Bacillus sp. F-33 was used as a representative of the Bacillus sp. strains with the other endophytic strains, Klebsiella sp. Sal 1, Enterobacter sp. Sal 3, and Herbaspirillum sp. Sal 6, to examine the effect of co-inoculation on their plant growth promotion and colonization in the tomato plant.
Each bacterial strain was cultivated under the same conditions as described in Section 2.2 to prepare the inoculum at ca. 10 8 CFU/mL. In case of co-inoculation, the same volume of individual cell suspension was mixed. The sterilized seeds were sown in the sterilized vermiculite in a capped glass tube (12 cm × 3 cm) supplied with the sterilized Hoagland solution, and 1 mL of the inoculum was added onto the seed zone. The other procedures were the same as those described in Section 2.2 except that the cultivation period was 14 days, and that the plant samples were macerated using a BioMasher (Nippi, Tokyo, Japan). The morphologies of the colonies of the co-inoculated strains were clearly different for counting separately. The inoculation experiment was conducted twice.

Effect of Time Interval Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 and Klebsiella sp. Sal 1 in Tomato Plant
Bacillus sp. F-33 and Klebsiella sp. Sal 1 were used as representatives of the Bacillus sp. and the endophytic strains, respectively, to examine the effect of time interval of inoculation on their plant growth promotion and colonization in the tomato plant. The experimental procedures were the same as those described in Section 2.3 except that Bacillus sp. F-33 was inoculated first, and then Klebsiella sp. Sal 1 was separately inoculated 7 days after the first inoculation. The tomato plants were harvested at 14 days after the first inoculation. An experiment with a different order of inoculation, Klebsiella sp. Sal 1 first and Bacillus sp. F-33 s, was also conducted in the same way. The inoculation experiment was conducted twice, but one experiment was done using two plants and one of the plants was used to determine the population.

Statistical Analysis
Statistical analysis of the data on the plant growth and population of the inoculant obtained in each twice-repeated experiment was performed using the MSTAT-C 6.1.4 [34] software package. Data were subjected to Tukey's test after one-way ANOVA.

Plant Growth Promotion and Colonization of Bacillus sp. Strains in Tomato Plant
The effects of inoculation of the Bacillus sp. strains on the growth of the tomato plant are presented in Figure 1. All Bacillus sp. strains except for Bacillus sp. RF-37 showed plant growth promotion. The root and shoot weights, and the shoot lengths of the inoculated tomato plant were significantly larger than the control while the root lengths were not affected. More lateral root development was observed in the inoculated tomato plant compared with the control.
after the first inoculation. The tomato plants were harvested at 14 days after the first inoculation. An experiment with a different order of inoculation, Klebsiella sp. Sal 1 first and Bacillus sp. F-33 s, was also conducted in the same way. The inoculation experiment was conducted twice, but one experiment was done using two plants and one of the plants was used to determine the population.

Statistical Analysis
Statistical analysis of the data on the plant growth and population of the inoculant obtained in each twice-repeated experiment was performed using the MSTAT-C 6.1.4 [34] software package. Data were subjected to Tukey's test after one-way ANOVA.

Plant Growth Promotion and Colonization of Bacillus sp. Strains in Tomato Plant
The effects of inoculation of the Bacillus sp. strains on the growth of the tomato plant are presented in Figure 1. All Bacillus sp. strains except for Bacillus sp. RF-37 showed plant growth promotion. The root and shoot weights, and the shoot lengths of the inoculated tomato plant were significantly larger than the control while the root lengths were not affected. More lateral root development was observed in the inoculated tomato plant compared with the control. The tomato plant was cultivated using sterilized vermiculite, and the parameters were measured at 24 days after seed inoculation. CTL represents the control samples inoculated with autoclaved cultures. The bars represent the standard deviation (n = 6), and different letters indicate significant differences at p < 0.05 by Tukey's test.
The populations of the inoculated Bacillus sp. strains in the rhizosphere, root, and shoot of the tomato plants are presented in Figure 2. All Bacillus sp. strains were detected in the rhizosphere, root, and shoot, and the populations of Bacillus sp. RF-12 and RF-37, which were originally isolated from the rhizosphere of sweet potato, and that of Bacillus sp. F-33, which was originally isolated as an endophyte of sweet potato, were higher than that of Bacillus sp. OYK, which was originally isolated from soil. The populations of the three Bacillus sp. strains were 0.9-2.2, 2.1-2.8, and 1.4-2.2 orders higher than those of Bacillus sp. OYK in the rhizosphere, root, and shoot, respectively. The populations were 2.4- Figure 1. The effects of inoculation of Bacillus sp. strains on the growth of the tomato plant. The tomato plant was cultivated using sterilized vermiculite, and the parameters were measured at 24 days after seed inoculation. CTL represents the control samples inoculated with autoclaved cultures. The bars represent the standard deviation (n = 6), and different letters indicate significant differences at p < 0.05 by Tukey's test.
The populations of the inoculated Bacillus sp. strains in the rhizosphere, root, and shoot of the tomato plants are presented in Figure 2. All Bacillus sp. strains were detected in the rhizosphere, root, and shoot, and the populations of Bacillus sp. RF-12 and RF-37, which were originally isolated from the rhizosphere of sweet potato, and that of Bacillus sp. F-33, which was originally isolated as an endophyte of sweet potato, were higher than that of Bacillus sp. OYK, which was originally isolated from soil. The populations of the three Bacillus sp. strains were 0.9-2.2, 2.1-2.8, and 1.4-2.2 orders higher than those of Bacillus sp. OYK in the rhizosphere, root, and shoot, respectively. The populations were 2.4-4.0 and 3.1-5.2 orders higher in the rhizosphere than those in the root and shoot, respectively. No colony appeared in the control samples.

Effect of Co-Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 with the Other Endophytic Strains in Tomato Plant
The effects of co-inoculation of Bacillus sp. F-33 with the other endophytic strains are presented in Figure 3. The plant growth tended to be promoted by Bacillus sp. F-33 but not significantly. The reduction tendencies of the effects were observed by co-inoculation of Enterobacter sp. Sal 3 and Herbaspirillum sp. Sal 6. In shoot weight and root length, the effects of the co-inoculation seemed to be negative in most cases.

Effect of Co-Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 with the Other Endophytic Strains in Tomato Plant
The effects of co-inoculation of Bacillus sp. F-33 with the other endophytic strains are presented in Figure 3. The plant growth tended to be promoted by Bacillus sp. F-33 but not significantly. The reduction tendencies of the effects were observed by co-inoculation of Enterobacter sp. Sal 3 and Herbaspirillum sp. Sal 6. In shoot weight and root length, the effects of the co-inoculation seemed to be negative in most cases.
Microbiol. Res. 2021, 12, FOR PEER REVIEW 5 4.0 and 3.1-5.2 orders higher in the rhizosphere than those in the root and shoot, respectively. No colony appeared in the control samples.

Effect of Co-Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 with the Other Endophytic Strains in Tomato Plant
The effects of co-inoculation of Bacillus sp. F-33 with the other endophytic strains are presented in Figure 3. The plant growth tended to be promoted by Bacillus sp. F-33 but not significantly. The reduction tendencies of the effects were observed by co-inoculation of Enterobacter sp. Sal 3 and Herbaspirillum sp. Sal 6. In shoot weight and root length, the effects of the co-inoculation seemed to be negative in most cases.  All strains colonized tomato plants, resulting in a large population, in which those of the endophytic strains were 1.5-1.7, 1.7-2.6, and 1.2-2.3 orders higher than those of Bacillus sp. F-33 in the rhizosphere, root, and shoot, respectively (Figure 4). Among the endophytic strains, the populations were not different in the rhizosphere, but the populations of Herbaspirillum sp. Sal 6 were about one order of magnitude higher than Klebsiella sp. Sal 1 in the plant parts. The populations were 1.8-2.7 and 2.3-3.3 orders higher at the rhizosphere than those in the root and shoot, respectively. No colony appeared in the control samples.
Microbiol. Res. 2021, 12, FOR PEER REVIEW 6 inoculated with autoclaved cultures. The bars represent the standard deviation (n = 6), and different letters indicate significant differences at p < 0.05 by Tukey's test.
All strains colonized tomato plants, resulting in a large population, in which those of the endophytic strains were 1.5-1.7, 1.7-2.6, and 1.2-2.3 orders higher than those of Bacillus sp. F-33 in the rhizosphere, root, and shoot, respectively (Figure 4). Among the endophytic strains, the populations were not different in the rhizosphere, but the populations of Herbaspirillum sp. Sal 6 were about one order of magnitude higher than Klebsiella sp. Sal 1 in the plant parts. The populations were 1.8-2.7 and 2.3-3.3 orders higher at the rhizosphere than those in the root and shoot, respectively. No colony appeared in the control samples. In case of the co-inoculation, no apparent change in the population was observed in most cases. In co-inoculation of Bacillus sp. F-33 and Herbaspirillum sp. Sal 6, however, the population in the shoot tended to decrease by 0.8 and 1.8 orders in Bacillus sp. F-33 and Herbaspirillum sp. Sal 6, respectively. In addition, one example of a positive tendency in the co-inoculation was observed in the population of Klebsiella sp. Sal 1 in the shoot, in which a 1.4-order increase was observed.

Effect of Time Interval Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 and Klebsiella sp. Sal 1 in Tomato Plant
The effects of the time interval of inoculation of Bacillus sp. F-33 and Klebsiella sp. Sal 1 are presented in Figure 5. The plant growth seemed to be promoted by Bacillus sp. F-33 but not by Klebsiella sp. Sal 1. When Klebsiella sp. Sal 1 was inoculated after Bacillus sp. F-33, the plant growth-promoting effects tended to be reduced in root weight. On the other hand, when Bacillus sp. F-33 was inoculated after Klebsiella sp. Sal 1, the effects seemed to be increased compared with the single inoculation of Klebsiella sp. Sal 1. In case of the co-inoculation, no apparent change in the population was observed in most cases. In co-inoculation of Bacillus sp. F-33 and Herbaspirillum sp. Sal 6, however, the population in the shoot tended to decrease by 0.8 and 1.8 orders in Bacillus sp. F-33 and Herbaspirillum sp. Sal 6, respectively. In addition, one example of a positive tendency in the co-inoculation was observed in the population of Klebsiella sp. Sal 1 in the shoot, in which a 1.4-order increase was observed.

Effect of Time Interval Inoculation on Plant Growth Promotion and Colonization of Bacillus sp. F-33 and Klebsiella sp. Sal 1 in Tomato Plant
The effects of the time interval of inoculation of Bacillus sp. F-33 and Klebsiella sp. Sal 1 are presented in Figure 5. The plant growth seemed to be promoted by Bacillus sp. F-33 but not by Klebsiella sp. Sal 1. When Klebsiella sp. Sal 1 was inoculated after Bacillus sp. F-33, the plant growth-promoting effects tended to be reduced in root weight. On the other hand, when Bacillus sp. F-33 was inoculated after Klebsiella sp. Sal 1, the effects seemed to be increased compared with the single inoculation of Klebsiella sp. Sal 1. In the time interval of inoculation, F-33 + Sal 1 and Sal 1 + F-33, the second inoculation was conducted 7 days after the first inoculation and analyzed 7 days after the second inoculation. CTL represents the control samples inoculated with autoclaved cultures. The bars represent the standard deviation (n = 5), and different letters indicate significant differences at p < 0.05 by Tukey's test.
In individual inoculation, populations of Klebsiella sp. Sal 1 were 1.9, 1.7, and 3.0 orders higher than those of Bacillus sp. F-33 in the rhizosphere, root, and shoot, respectively, and the populations were 2.7-2.8 and 2.5-3.7 orders higher in the rhizosphere than those in the root and shoot, respectively ( Figure 6). When Klebsiella sp. Sal 1 was inoculated after Bacillus sp. F-33, the populations of Bacillus sp. F-33 were similar to those in the individual inoculation. When Bacillus sp. F-33 was inoculated after Klebsiella sp. Sal 1, those were 1.3-2.4 orders lower than those in individual inoculation. The populations of Klebsiella sp. Sal 1 showed similar levels under any conditions. No colony appeared in the control samples.  The tomato plant was cultivated using sterilized vermiculite, and the parameters were measured at 14 days after seed inoculation. In the time interval of inoculation, F-33 + Sal 1 and Sal 1 + F-33, the second inoculation was conducted 7 days after the first inoculation and analyzed 7 days after the second inoculation. CTL represents the control samples inoculated with autoclaved cultures. The bars represent the standard deviation (n = 5), and different letters indicate significant differences at p < 0.05 by Tukey's test.
In individual inoculation, populations of Klebsiella sp. Sal 1 were 1.9, 1.7, and 3.0 orders higher than those of Bacillus sp. F-33 in the rhizosphere, root, and shoot, respectively, and the populations were 2.7-2.8 and 2.5-3.7 orders higher in the rhizosphere than those in the root and shoot, respectively ( Figure 6). When Klebsiella sp. Sal 1 was inoculated after Bacillus sp. F-33, the populations of Bacillus sp. F-33 were similar to those in the individual inoculation. When Bacillus sp. F-33 was inoculated after Klebsiella sp. Sal 1, those were 1.3-2.4 orders lower than those in individual inoculation. The populations of Klebsiella sp. Sal 1 showed similar levels under any conditions. No colony appeared in the control samples.
ders higher than those of Bacillus sp. F-33 in the rhizosphere, root, and shoot, respectively, and the populations were 2.7-2.8 and 2.5-3.7 orders higher in the rhizosphere than those in the root and shoot, respectively ( Figure 6). When Klebsiella sp. Sal 1 was inoculated after Bacillus sp. F-33, the populations of Bacillus sp. F-33 were similar to those in the individual inoculation. When Bacillus sp. F-33 was inoculated after Klebsiella sp. Sal 1, those were 1.3-2.4 orders lower than those in individual inoculation. The populations of Klebsiella sp. Sal 1 showed similar levels under any conditions. No colony appeared in the control samples.

Discussion
Significant plant growth-promoting properties were observed in the Bacillus sp. strains except for Bacillus sp. RF-37 ( Figure 1). Similar PGPR properties in Bacillus spp. have been previously reported [35][36][37][38]. In this study, the inoculants stimulated lateral root growth, resulting in greater root weight, which could explain the inconsistent results on root weight and root length in the inoculated plants. As indole-3-acetic acid (IAA) is known to have similar effects on plants [39], the plant growth promotion might be caused by IAA production by the inoculants. In another experiment, Bacillus sp. RF-12 and F-33 showed an IAA-producing ability while Bacillus sp. RF-37 did not (data not shown). However, since Bacillus sp. OYK also showed no activity, the reason for the plant growth promotion is unclear.
In our previous study, the inoculated Bacillus sp. OYK strain could not establish its population as an endophyte in sweet potato [14], although Bacillus spp. strains have been reported as indigenous endophytes in sweet potato [15,16], tomato [19], banana [20], and switchgrass [21]. We attributed it to the competition with indigenous rhizobacteria and endophytes, as well as the endophytic ability of the inoculant.
In this study, all Bacillus strains colonized in the rhizosphere and endosphere of the tomato plants cultivated using sterilized vermiculite (Figure 2), suggesting that Bacillus sp. OYK has endophytic potential, and that the presence of indigenous microorganisms inhibited its colonization. However, the 1.4-2.8-orders lower populations of Bacillus sp. OYK in the plants compared with the other Bacillus sp. strains, which were isolated from the rhizosphere or as an endophyte (Figure 2), suggests decreased competitiveness of Bacillus sp. OYK against indigenous plant-associated microbes. Some genes and functions may be involved in the plant colonization ability, and PGPR strains from different habitats may have different interactions with plants. The use of originally plant associated PGPR could establish their populations at the rhizosphere and/or endosphere of plants.
The plant growth-promoting effects of Bacillus sp. F-33 were reduced in the presence of the other endophytes, though the population of Bacillus sp. F-33 was maintained (Klebsiella sp. Sal 1 and Enterobacter sp. Sal 3) or slightly decreased (Herbaspirillum sp. Sal 6) (Figures 3 and 4), suggesting that its phyto-stimulating ability was neutralized by the other strains. As the three co-inoculated strains have IAA-degrading ability [30], they might degrade IAA produced by Bacillus sp. F-33 below the effective level.
Synergetic effects of co-inoculation have been reported [23][24][25][26] while cancelation of the positive effects [43][44][45], and negative effects of co-inoculation have also been reported [29,30]. The effects of the co-inoculation seemed to be dependent on the combination of the strains. In most studies that examined the effects of co-inoculation of PGPR, changes in populations of the PGPR by co-inoculation were not measured. In the limited examples of the study using Azospirillum brasilense Sp245 and Bacillus subtilis 101 [29], and Klebsiella sp. Sal 1 and Herbaspirillum sp. Sal 6 [30], their plant growth promotions were reduced even though the populations of the PGPR were maintained, as observed in this study. In our previous study, diverse endophytic bacterial communities were observed in sweet potato, and some components of the communities disappeared by inoculation of Bacillus sp. OYK [14]. It is crucial to elucidate the mechanisms of the microbial interactions; however, it might be complex given the actual environment.
After the establishment of Bacillus sp. F-33 in the rhizosphere and in the tomato plant, Klebsiella sp. Sal 1 could colonize the same population as the strain was individually inoculated ( Figure 6) and inhibited the plant growth-promoting ability of Bacillus sp. F-33 without reducing its population ( Figure 5), as in the co-inoculation experiment. The high colonizing potential of Klebsiella sp. Sal 1 seemed not to be affected by the about 2-orders lower population of the previously established Bacillus sp. F-33.
On the other hand, after the establishment of Klebsiella sp. Sal 1, the colonization of Bacillus sp. F-33 was reduced by 1.3-2.4 orders than those in the individual inoculation ( Figure 6). The relatively lower potential for colonization of Bacillus sp. F-33 might be the reason. The microbial community structure might be a crucial factor to determine the fate of allochthonous microorganisms, such as a PGPR inoculant. Pre-inoculation of PGPR prior to transplantation could be one practical method to enhance higher colonization in plants.
In spite of the reduced population of Bacillus sp. F-33, the plant growth promotion was increased when the strain was inoculated after Klebsiella sp. Sal 1 ( Figure 5). It was suggested that the level of the population is not a determinant of the potential of the strain. Although the population of Bacillus sp. F-33 was maintained both in the co-inoculation and in the inoculation of Klebsiella sp. Sal 1 after Bacillus sp. F-33, the PGPR potential of Bacillus sp. F-33 was reduced in the presence of Klebsiella sp. Sal 1, so unknown factors might be involved in plant growth promotion. In addition, the ratio between the populations might not be constant when plants developed, and the kinetic of the different bacterial populations might not be reflected by one sampling time. Time course analysis after inoculation could reveal the progress of colonization in the plant. The results of this study also indicate that there are different niches for the different strains and the colonization of these niches may not have the same impact on plant growth. It may mean that bacteria are competing for some niche colonization.
In addition to plant growth-promoting properties, the colonization potential should be considered as important criteria when assessing their suitability for commercial development. The lower population of Bacillus sp. OYK, which was isolated from soil, than the other Bacillus sp. strains, which were isolated from either the rhizosphere or endosphere of plant samples, suggests the importance of the origin of the strains for their colonization. The plant growth promotion and colonization potentials were independently affected by the co-existing microorganisms. Further studies are necessary to evaluate the colonization potential of PGPR under field conditions where diverse microorganisms exist.

Conclusions
In this study, the higher population of rhizospheric and endophytic Bacillus sp. in the plant suggest the importance of the origin of the strains for their colonization. The plant growth promotion and colonization potentials were independently affected by the coexisting microorganisms.