*3.4. High-Throughput DNA Sequencing of LFM*

In total, 192,355 reads were obtained; the average number of observed species per sample was 1013 ± 170 (max: 1306, min: 892), and the coefficient of variation was 0.17. The average bacterial composition was shown via the integration of the clustering results and the relative abundance of each sample by phylum (Figures 1a,b and S3). *Proteobacteria* were most frequently detected, followed by *Firmicutes* and *Actinobacteria*. The species composition at the phylum level was different in August when compared with that from the other months (Figure 1a). At the genus level, when the top 10 genera were compared between different months, the genera composition of the LFM in May was different from that from the other months (Figures 1a,b and S3). This is because there were few *Lactobacillus* spp. at this time, and the fermentation was in the early stages.

**Figure 1.** (**a**) Unweighted pair–group method with arithmetic means cluster tree based on unweighted and weighted unifrac distance. This was displayed with the integration of clustering results and the relative abundance of each sample by phylum. (**b**) The top 10 taxa at the genus level were selected to form the distribution histogram of relative abundance.

#### *3.5. Incubation Study of LFM Utilization*

The release of NO3 <sup>−</sup>-N and phosphate was used to determine the potential for nutrient provision from LFM. The release of NO3 <sup>−</sup>-N started from 14 days in the LFM, and approximately 200 mg/kg−<sup>1</sup> of NO3 <sup>−</sup>-N had accumulated by the end of 14 weeks of incubation (Figure S1a). Phosphate availability followed the same trend as that of NO3 <sup>−</sup>-N release (Figure S1b). Available phosphate was rapidly released from rapeseed cake, whereas no phosphate was detected from LFM until week 2. From week 3, the availability of phosphates increased slightly in the LFM.

### *3.6. Pot Experiments*

A pot experiment was conducted to examine the effect of selected strains on the growth of brassica and lettuce. In the initial experiment on lettuce, 11 strains were tested; 4 of these strains showed high activity when compared with the control. Therefore, strains 2, 4, 6 and 11 were tested on brassica, and showed significant differences in their growth-promoting effect (Figure 2). A pot experiment was also conducted to assess the effect of LFM on the growth of both brassica and lettuce. There was no significant difference in the growth characteristics of lettuce between the LFM and chemical fertilizer treatments (Figure 3a); however, brassica exhibited significant differences in its root biomass with LFM amendment (Figure 3b, Tukey's HSD, *p* < 0.05).

**Figure 2.** Growth response of *Brassica campestris* with plant growth-promoting bacteria selected in this study. Control (*n* = 7), strain number 2 (*n* = 10), strain number 4 (*n* = 10), strain number 6 (*n* = 7), strain number 11 (*n* = 10), Dunnett test (*p* < 0.05). The vertical bar indicates the standard error. \* indicates significance differences between treatments when compared with the control (*p* < 0.05).

**Figure 3.** Pot experiment testing lettuce (**a**) and *Brassica rapa* (**b**) growth using liquid food waste materials (LFM). Values presented are means and standard error (*n* = 3). Closed bar and Gray bar mean edible part and root, respectively. Treatments of the same crop with different letters are significantly different by Tukey's HSD (*p* < 0.05).

#### *3.7. Field Experiments*

Field experiments were conducted to assess the effect of LFM on the growth of *Brassica rapa*, lettuce and eggplant. The growth of *Brassica rapa* and lettuce in the field was similar to that in the pot experiment. The heights achieved by *Brassica rapa* were 33.0 ± 0.67 cm and 32.5 ± 0.78 cm, with LFM and chemical fertilizer, respectively. The lettuce grown with LFM amendment was slightly larger than that grown with chemical fertilizer, but not significantly so (Figure 4a). Eggplant also grew slightly better with LFM than with chemical fertilizer, but the differences were not significant (Figure 4b,c for eggplant).

**Figure 4.** Field experiment testing the effect of liquid food waste materials (LFM) on the growth of Lettuce (**a**) and eggplant (**b**,**c**).

#### **4. Discussion**

This study showed that useful microorganisms, such as PGPB, were present in LFM produced from the recycling of unused resources. This is an important finding that leads to the promotion of recycling, and also indicates that unused resources are useful as microbial resources. LFM can be used as a fertilizer, and has other positive effects on vegetable growth. However, there are limitations to the plant growth-enhancement functions of LFM. If we apply isolated strains to LFM in order to promote a plant growth function more effective than that of the original LFM, some supplementary nutrients are needed to produce an effect. This point still remains an issue.

High-throughput sequencing analysis of the bacterial community composition in the LFM revealed substantial differences between the sampling months May and August. A similar bacterial community composition was observed for the samples from June, July and September. The May sample was still in the early stages of fermentation, and there was no school-provided lunch in August; therefore, LFM was stored in the tank, which might help to explain the differences in the community composition of these samples.

Most of the bacteria isolated from LFM were *Bacillus* spp., which is a spore-forming bacteria. This was because the LFM pH was reduced to 3.0 through the process of lactic acid fermentation, and the temperature exceeded 50 ◦C. However, this result was obtained because we used R2A media for isolation, and we detected the family of *Lactobacillaceae* in the LFM through high-throughput sequencing (Figure S3). The most frequently isolated *Bacillus* spp. strains were *Bacillus amyloliquefaciens* (56.7%) and *Bacillus vallismortis* (26.7%). *Bacillus* species are known to produce dormant spores [44], and enact an anti-pathogen activity through the assembly of non-ribosomal cyclic lipopeptides [45]. In addition, *Bacillus* species are considered PGPB because of their potential for antibiotic production, biofilm formation on the plant root surface, and production of plant hormones [46,47]. Furthermore, seed treatment with *Bacillus* species has been shown to significantly enhance shoot fresh and dry weight, as well as plant height, in various crops [48,49]. From the results of this study, >1000 bacterial OTUs were identified in LFM; therefore, it might be possible to isolate other useful strains, other than *Bacillus* spp., under a range of isolation conditions, including increased pH.

Organic materials slowly release nutrients, but they are still a promising alternative to chemical fertilizers, as their application can reduce nutrient leaching, volatilization and problems of toxicity [50]. In the present study, LFM was used as an alternative to chemical fertilizers in order to investigate the release of NO3 <sup>−</sup>-N and available phosphate (Figure S1). Low amounts of available phosphate were released from LFM during the incubation study because of the low total phosphate concentration in LFM. LFM released NO3 <sup>−</sup>-N from the third week of incubation, and released approximately 200 kg/ha−<sup>1</sup> NO3 <sup>−</sup>-N during the 13 weeks of incubation. Moreover, the biomass richness of soil fertilized with LFM was higher than that treated with chemical fertilizer [51], and LFM did not change the soil pH after treatment through our study.

The growth-promoting effect of strain number 11, with an IAA-producing ability and an ACC deaminase activity, was confirmed in a pot experiment; growth was significantly promoted with inoculation by strain number 11, when compared with the control (Figure 2). The growth of lettuce in a pot was the same with both LFM and chemical fertilizer. For *Brassica rapa*, the growth of the edible (aboveground) part was the same with both LFM and chemical fertilizer, whereas the root biomass was significantly increased with LFM, when compared with chemical fertilizer (Figure 3b). These results indicate that the PGPB in LFM contributed to the increase in root biomass of *Brassica rapa*. A field experiment was conducted to assess the effect of LFM on the growth of *Brassica rapa*, lettuce and eggplant. Although the N input by LFM was less than half that of the chemical fertilizer, the growth of the eggplant with LFM was the same as that with chemical fertilizer. The yield with LFM was higher than, but not significantly different from, that achieved with chemical fertilizer. These results indicate the abundance and activity of PGPB in LFM, and their efficacy in supporting eggplant growth under the conditions tested. LFM could be a viable alternative to commercially available chemical fertilizers, without an adverse effect on soil and vegetable growth. A previous study showed the positive effect of PGPB inoculation on vegetable growth and yield [52].

All the selected strains that showed growth-enhancement in the pot experiment had an IAA-producing ability (Table 1). IAA is a type of plant hormone that promotes root elongation and enhances root growth. Many PGPBs with the IAA-production ability have been isolated in previous studies [1,53]. Furthermore, all strains that were positive for IAA production also showed ACC deaminase activity (strain numbers 2, 4, 6 and 11). This suggests that IAA production and ACC deaminase activity contribute greatly to enhancing the plant growth in our isolated strains, while *Caulobacter* sp. had a negative impact on plant growth, even though it produced higher levels of IAA [54]. In addition, previous studies have shown that PGP microbes and PGPB can promote plant growth indirectly or directly, through the production of ACC deaminase and through reducing the ethylene level in the developing plants through the roots [52,55], by generating plant growth hormones like IAA [56]. It is likely that ACC deaminase and IAA production promote root growth in a similar fashion [57,58]. Of the selected strains, only strain number 11 showed phosphorus solubilizing potential and siderophore production (Table 1). Phosphate solubilization is effective in soils with low available phosphoric acid, and siderophore production chelates the iron in soils with high pH to help plant uptake [52,59]. However, the detailed mechanism of plant growth-enhancement is complex, and further investigation is needed [54,60].

The selection of PGPB strains from LFM was important to confirm the positive effect of the inoculants on plant growth, and to optimize their application for maximum impact on vegetable crops. The main aim of this study was to reduce the commercial use of chemical fertilizers, by utilizing LFM as an alternative fertilizer with the maximum impact on crop growth and soil, and minimal environmental impact. Further investigation into LFM use as an organic fertilizer should evaluate any adverse impact of its application to the soil environment.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4395/10/7/954/s1, Figure S1: Nitrate nitrogen (**a**) and available phosphate (**b**) release from organic materials, • rapeseed oil cake, - Liquid food waste material (LFM). Figure S2: 31 strains isolated from liquid food waste material belong to genus *Bacillus*. Type A closely related to *Bacillus velezensis* strain FZB42, type B closely related to *Bacillus amyloliquefaciens* strain MPA 1034, type C closely related to *Bacillus vallismortis* strain NRRL B-14890, type D closely related to *Bacillus subtilis* subsp. inaquosorum strain BGSC 3A28, type E closely related to *Bacillus wiedmannii* strain FSL W8-0169, type F closely related to *Bacillus velezensis* strain NTGB-29, and type G closely related to *Bacillus vallismortis* strain DSM 11031. Figure S3: Top 30 bacterial compositions, in different taxonomic levels.

**Author Contributions:** Conceptualization, R.K.; methodology, S.K., R.I., R.K., W.A., A.M.; writing—original draft preparation, R.K., W.A., A.M.; writing—review and editing, R.K.; visualization, R.K., W.A.; supervision, R.K.; project administration, R.K.; funding acquisition, R.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Strategic research project in university of yamanashi, grant number 14 and the Japan Society for the Promotion of Science (JSPS), Ministry of Education, Culture, Sports, Science and Technology, Japan [KAKENHI Grant Number 19K12379].

**Acknowledgments:** The authors are grateful to Mr. Yuta KOBAYASHI for supporting to manage the field conditions. We also thank Kai city for their many supports.

**Conflicts of Interest:** The authors declare no conflict of interest.
