The study was divided into three phases (Table S1). First, we examined potential carriers for the microbial communities and selected the most suitable carrier for subsequent experiments. Second, we tested a variety of conditions (temperature, amount of organic material, and average RH during incubation) to determine the optimal conditions for nitrification. Third, we determined the microbial taxa that were present in the medium and the effects of the optimal conditions on the growth of several crop plants.
Selection of carriers
We tested 12 soil-less substrates: rockwool (grain size 3 to 7 mm; Grodan, Roermond, Netherlands; bulk density [ρ] 219.7 g/L); vermiculite (Nitto Hiruishi, Ianabe, Japan; ρ 147.6 g/L); pumice (Setogahara Kaen, Midori, Japan; ρ 408.7 g/L); perlite (Katanagawa Heiwa Nouen, Kanuma, Japan; ρ 60.0 g/L); coconut husk (Yoshimoto Nousan, Nangoku, Japan; ρ 78.0 g/L); 5 polyurethane resins (7.5 mm, Inoac Corporation, Nagoya, Japan: AQ-20, ρ 22.7 g/L; AQ-15, ρ 18.5 g/L; AQ-14, ρ 21.3 g/L; Mixel GR, ρ 41.0 g/L; Mixel GP, ρ 32.0 g/L); oyster shell lime (OSL; Suzuki Yuukan Center, Miyagi, Japan; ρ 693.8 g/L); and rice husk charcoal (Komeri, Niigata, Japan; ρ 110.2 g/L). Table S2 summarizes the physical characteristics of the substrates. We packed 100 mL of each material into an open-bottom tube made from the inverted top half of a 350-mL plastic soft-drink bottle (EBM, Tsubame, Japan), added 1 g of powdered bark compost (Sanyo bark, Sanyo Chip Kogyo, Shimonoseki, Japan) as the microbial inoculum, and then rinsed the materials with 100 mL of distilled water. We added 1 mL of 10% w/v fish fertilizer diluted with distilled water, and then incubated the tubes for 24 h in the dark at 25 °C in an incubator (CN-25C; Mitsubishi Electric, Tokyo, Japan). The next day, the carriers were rinsed with 100 mL of distilled water and the concentrations of ammonium, nitrite, and nitrate in the leachate were measured. The operation of adding fish fertilizer, incubating overnight at 25 °C, and rinsing with water was repeated daily for 2 weeks. Three tubes were used per experiment.
Determination of optimal conditions
To immobilize microorganisms on the carriers16, we packed a weight equivalent to 100 mL of rockwool carrier (22 g; grain size 3 to7 mm; Grodan, ρ 219.7 g/L) in a tube and added 1 g of powdered bark compost (Sanyo bark), as inoculum (Fig. S1). We then rinsed the carriers with 100 mL of distilled water, added 1 mL of 10% w/v fish-based soluble fertilizer (Yaizu Suisankagaku Industry, Yaizu, Japan) diluted with distilled water, and incubated the tubes for 24 h in the dark at 25 °C in the CN-25C incubator. The next day, the carriers were rinsed with 100 mL of distilled water and the concentrations of ammonium, nitrite, and nitrate in the leachate were measured. The operation of adding fish fertilizer, incubating overnight at 25 °C, and rinsing with water was repeated daily for 2 to 3 weeks until nitrate was detected in the leachate. After a microbial community was established by completing this operation, the rockwool was used as a microorganism carrier in subsequent experiments. The chemical properties of the leachates from the carrier were measured daily for 30 days. Rockwool that was not inoculated with microorganisms but that was otherwise treated in the same manner was used as the control in subsequent experiments. These experiments were each performed in three tubes.
The following experiments were performed using tubes of rockwool carrier in which the addition of 6 mg N of fish fertilizer, overnight incubation at 25 °C, and rinsing with 100 mL of distilled water were repeated daily for 2 weeks until nitrate was detected in the leachate. After the detection of nitrate, we repeated the addition of fish fertilizer, overnight incubation, and rinsing with water daily for an additional 2 weeks in each experiment. All experiments were each performed in three tubes. To investigate the effect of excess organic substances on the production of inorganic N, we added 1 to 10 mL of 10% w/v fish fertilizer diluted with distilled water per tube daily so as to add 6 to 60 mg organic N. After incubation overnight, each tube was rinsed with 100 mL of distilled water daily.
In the subsequent experiments, we added 6 mg N of fish fertilizer (which provided the optimal results in the previous stage of the experiment). To determine the optimal incubation temperature, tubes were incubated at 15, 20, 25, 30, 37, 42, or 45 °C. To determine the optimal RH, tubes with materials that exhibited nitrification were incubated in incubators at an RH of 20% (WFO-600ND, Eyela, Tokyo, Japan), 51% (MLR-352, Panasonic, Kadoma, Japan), or 92% (CN-25C) at 25 °C. The amounts of inorganic N, ammonium, nitrite, and nitrate in the leachates were measured under each experimental condition.
The following experiments were performed using tubes of rockwool carrier in which the addition of fish fertilizer, overnight incubation, and rinsing with water had been repeated daily for 2 to 3 weeks until nitrate was detected in the leachate. To compare the effects of different organic substances, we added fish fertilizer (6% N w/w, C/N ratio 2.9), corn steep liquor (CSL, OAT Agrio, Tokyo, Japan; 3.3% N w/w, C/N ratio 4.8), or rapeseed oil cake (Sun and Hope, Kitakyushu, Japan; 6% N w/w, C/N ratio 6.9) as sources of organic N in amounts equivalent to 6 mg N per tube and then incubated the tubes overnight in the dark at 25 °C. The tubes were then rinsed with 100 mL of distilled water, followed by the addition of organic substances and incubation, daily for 3 weeks. The amounts of inorganic N, ammonium, nitrite, and nitrate in the leachates were then measured. These experiments were each performed in three tubes.
Identification of microbes and confirmation of plant growth
To estimate the microbial density, we added 1 g (fresh weight) of the sample carrier (rockwool) to 9 mL of sterilized water and vortexed the mixture for 5 min; we then applied 100 μL of the sample to 1/10 NA medium (0.8 g/L Difco nutrient broth [Becton Dickinson, Franklin Lakes, NJ, USA] + 15 g/L agar [Fujifilm Wako, Osaka, Japan]) using the dilution plate technique. We incubated the plates at 25 °C for 1 week in the dark and counted the colonies. To detect and count the nitrifying bacteria, we used two antibody-based determination kits (Kenshutsu-kun and Spira-kun; Yakult, Tokyo, Japan).
We analysed the microbial phase in the rockwool carrier17. We prepared tubes filled with 100 mL of rockwool by rinsing the material with 100 mL of distilled water; we added 1 g of bark compost on the top as inoculum and 1 mL of 10% w/v fish fertilizer diluted with distilled water as the organic substance, and then incubated the tubes in the dark overnight at 25 °C. The next day, the tubes were rinsed with 100 mL of distilled water and then 1 mL of 10% w/v fish fertilizer was again added. The addition of fish fertilizer, incubation overnight, and rinsing with water was repeated daily for 3 weeks until nitrate was detected in the leachate. Granules of rockwool that carried a microbial community were sent to TechnoSuruga Laboratory Co., Ltd. (Shizuoka, Japan) for amplicon sequence analysis17. Total DNA was extracted using the ISOIL for Beads Beating kit (Nippon Gene Co., Ltd., Toyama, Japan). Total DNA was purified with the DNeasy PowerClean Pro Cleanup Kit (Qiagen, Hulsterweg, The Netherlands). The V3–V4 regions of bacterial and archaeal 16S rRNA were amplified using the Pro 341F/805R primers and dual-index method18,19.
Identification from sequences
Sequence reads were analysed manually using the Ribosomal Database Project (RDP) Multiclassifier tool, which is available from the RDP website (http://rdp.cme.msu.edu/classifier/)20. Bacterial species were identified from sequences in Metagenome@KIN v. 2.2.1 analysis software (World Fusion, Japan) and in the TechnoSuruga Lab Microbial Identification database DB-BA 13.0 (TechnoSuruga Laboratory, Shizuoka, Japan) with homology for ≥ 97%21. We also analysed the microbial phase of the bark compost and the MPM liquid culture solution to provide a comparison. To prepare the MPM liquid culture solution, we placed 100 mL of sterilized distilled water with 1 g of bark compost as inoculum and added fish fertilizer to a final concentration of 1 g/L in a flask and cultured the solution on an orbital shaker (120 rpm) for 2 weeks at 25 °C in the dark. The culture solution was then centrifuged at 10,000×g for 5 min. The cell pellet was analysed by TechnoSuruga as described earlier in this section.
On rockwool that carried a microbial community, we sowed 5 seeds of komatsuna (Brassica rapa var. perviridis, ATU121 ‘Misaki’; Sakata Seed Corporation, Yokohama, Kanagawa, Japan), then added 0.1 g of fish fertilizer to each carrier and incubated the carriers in an MLR-352 chamber at 25 °C under a 12-h light/dark cycle for 24 h, and then rinsed the carrier with 100 mL of distilled water. We repeated the addition of fish fertilizer, incubation, and rinsing daily for 11 days. We also grew saladana lettuce (Lactuca sativa var. capitata ‘Santa Clara’; Tohoku Seed Co. Ltd., Utsunomiya, Tochigi, Japan), radish (Raphanus sativus var. sativus, ‘Akamaru Hatsuka Daikon’; Nihon Nousan Shubyo Co. Ltd., Kamiina, Nagano, Japan), turnip (Brassica rapa var. glabra, ‘Shogoin Kabu’; Atariya Nouen, Katori, Chiba, Japan), and ball lettuce (Lactuca sativa ‘Melbourne-MT’; Tohoku Seed Co. Ltd.) on rockwool and nursery soil (Naeichiban; Sumirin Agro-Products, Aichi, Japan) in a greenhouse of the Institute of Vegetable and Floriculture Science in Tsu, Mie Prefecture, from September to December 2013. During this period, photoperiod ranged from 305.9 to 371.5 h/month, with daily average outside temperatures of 6.0 to 23.5 °C, daily minimum outside temperatures of 1.7 to 19.4 °C, and daily maximum outside temperatures of 10.9 to 28.3 °C. Light and temperature were not controlled. Three pots per treatment were examined. We conducted these experiments in accordance with relevant institutional, national, and international guidelines and legislation. All plant seeds used were commercially available.
Methods for leachate analysis
We used Reflectoquant tests for the ammonium (Merck, Darmstadt, Germany; catalogue number 116892), nitrite (116973), nitrate (116971), phosphate (116978) and potassium (117945) measurements14. We added and mixed the supplied reagents according to the manufacturer’s instructions for each kit, submerged the test strip, and measured the degree of colour development using a reflective photometer (RQflex plus; Merck)14. When the leachate contained nitrite, we added 150 μL of 10% (w/v) amidosulfuric acid (Kanto Chemical, Tokyo, Japan) to 5 mL of the leachate before measuring nitrate, and mixed the solution to eliminate the influence of nitrite. pH and electrical conductivity (EC) were measured using a pH meter (C-73, AZ-ONE, Osaka, Japan) and an EC meter (Twin Cond, Horiba, Kyoto, Japan), respectively.
