We used two arable soils and two forest soils in this study to account for the effects of phosphorus fertilization history and land-use type. Arable soil no. 1 was collected from the Ap horizon at the Nagano Prefecture Vegetable and Ornamental Crop Experiment Station in Shiojiri, Nagano Prefecture, Japan. Arable soil no. 2 was obtained from the Ap horizon at the NARO Agricultural Research Center for the Hokkaido Region in Sapporo, Hokkaido Prefecture. Both soils were classified as Andisols according to the United States Department of Agriculture (USDA) soil taxonomy. Soil no. 2 was collected from a field that had been left fallow for several years without application of lime or fertilizers. Forest soils 3 and 4 were Inceptisols collected from the A horizon in Minamiaiki and Terasawa, respectively, in Nagano Prefecture. The soil samples were sieved through a 2-mm mesh and stored at 4 °C. A portion of the soil sample was also air-dried for chemical analyses.

Soil pH was measured in a 1:2.5 soil:solution suspension with distilled water or 1 M KCl. The total C and N were determined using an NC analyzer (JM1000CN, J-Science Lab, Kyoto, Japan). Potentially available P was estimated by the Truog method (9), with Truog-P extracted from a suspension with a 1:200 soil: solution ratio using 0.001 M H₂SO₄ buffered with (NH₄)₂SO₄ to a pH of 3.0. All concentrations were expressed on a dry weight basis and are shown in Table 1.

Bacterial and fungal contributions to phosphatases production in soils were evaluated using the selective inhibition methods developed by Hayano and Tubaki (10), Watanabe and Hayano (11), and Kunito et al. (12), with slight modifications. Briefly, the soil was oven dried at 105 °C for 12 h, after which antibiotics in aqueous solution were added to the soil (15 g on a dry weight basis) as follows (1): no addition (2), cycloheximide (2 mg g⁻¹ soil) + nystatin (2 mg g⁻¹ soil) to suppress fungal growth (3), chloramphenicol (1 mg g⁻¹ soil) + streptomycin (1 mg g⁻¹ soil) to suppress bacterial growth, or (4) cycloheximide (2 mg g⁻¹ soil) + nystatin (2 mg g⁻¹ soil) + chloramphenicol (1 mg g⁻¹ soil) + streptomycin (1 mg g⁻¹ soil) to suppress both fungal and bacterial growth. We determined the types and amounts of antibiotics applied in our preliminary experiment. Distilled water was added to the soil to 50% of the water holding capacity, and a small amount of untreated moist soil (0.2 g) was inoculated for each treatment. After a 7-day incubation at 23 °C, the activities of phosphatases and culturable microbial populations were determined. All samples were incubated in triplicate.

The activities of two phosphomonoesterases (acid phosphatase (EC 3.1.3.2) and alkaline phosphatase (EC 3.1.3.1)) and phosphodiesterase (EC 3.1.4.1) in soils were measured. Acid and alkaline phosphatase activities were measured using p-nitrophenyl phosphate as the substrate in a modified universal buffer (pH 6.5 and 11, respectively) as described by Tabatabai (13). Phosphodiesterase activity was measured using bis-p-nitrophenyl phosphate in tris (hydroxymethyl)aminomethane buffer (pH 8.0) (13). It should be noted that the measured activity represents potential rather than in-situ activity and reflects the amount of active enzyme present in the soil under substrate-saturating assay conditions (14). Also, because p-nitrophenyl phosphate do not cross the cytoplasmic membrane (15), the activities are likely derived extracellular enzymes.

In this study, we considered the possibility that a large amount of DNA derived from microbial residues might remain in the soil after heating treatment at 105 °C. Therefore, we evaluated microbial abundance by plate-count technique instead of quantitative PCR. Culturable bacteria were enumerated by a dilution plate-count technique using a diluted TSB agar plate (tryptic soy broth, 2.0 g; agar, 10 g; cycloheximide, 50 mg per liter) (16). Culturable fungi were enumerated using Martin’s rose Bengal streptomycin glucose agar plates (KH₂PO₄, 1.0 g; MgSO₄·7H₂O, 0.5 g; peptone, 5.0 g; glucose, 10 g; rose Bengal, 33 mg; agar, 20 g; streptomycin, 30 mg per liter) (17). Soil suspensions in sterile tap water (1:10 w/v) were dispersed with a Warring blender (Sakuma Seisakusyo, Japan) at 10,000 rpm for 3 min (18, 19), and the resulting slurry was decimally diluted with sterile tap water.

Tukey’s HSD test, along with one-way ANOVA, were conducted to compare the effects of the antibiotics on soil phosphatase activities and microbial counts. Goodness of fit to a normal distribution and homogeneity of variances were confirmed using the Shapiro–Wilk test and Bartlett’s test, respectively.

We used two arable soils and two forest soils in this study. Forest soils had higher organic matter content than arable soils (Table 1). While the addition of cycloheximide + nystatin markedly decreased culturable fungal abundance to near the detection limit (p <0.05), it had only a minor effect on culturable bacterial populations, resulting in bacteria-dominated soil samples (Figure 1). In contrast, the addition of chloramphenicol + streptomycin significantly reduced culturable bacterial abundance (p <0.05) without affecting fungal abundance in all but soil no. 4, leading to fungi-dominated soil samples in all other soils. These findings indicate that selective inhibition using antibiotics successfully established both bacteria- and fungi-dominated soil samples in this study. The inability to obtain soil samples consisting exclusively of bacteria or fungi is likely due to the ineffectiveness of the added antibiotics against microorganisms located within soil aggregates. It should be noted that the chloramphenicol + streptomycin treatment failed to significantly reduce culturable bacterial abundance in soil no. 4. However, the activities of phosphatases were low in this treatment; thus, the conclusion obtained was not affected. This is discussed in greater detail below. It is also noteworthy that chloramphenicol + streptomycin significantly increased fungi counts, whereas cycloheximide + nystatin significantly increased bacterial counts in some soils. These results are likely explained by the surviving microbial groups proliferating by utilizing the generated microbial necromass as substrates. For example, treatment with chloramphenicol + streptomycin likely caused substantial bacterial death, providing substrates that allowed fungi to grow.

The activities of all three phosphatases (except alkaline phosphatase in soil no. 2) were significantly higher in bacteria-dominated soil samples treated with cycloheximide + nystatin than in fungi-dominated soil samples treated with chloramphenicol + streptomycin (p <0.05) in soils 1, 2, and 4. Additionally, the activities of these phosphatases were comparable to those of the control without added antibiotics for these three soils (Figures 2–4). These results suggest that bacteria mainly produced all three phosphatases in soils 1, 2, and 4. In contrast, the activities of all three phosphatases were significantly higher in fungi-dominated soil samples treated with chloramphenicol + streptomycin than in bacteria-dominated soil samples treated with cycloheximide + nystatin (p <0.05) in soil no. 3, while they were comparable to those of the control without added antibiotics. This finding suggests that fungi mainly produced all three phosphatases in soil no. 3. A possible explanation for the dominant contribution of fungi to phosphatases production observed only in soil no. 3 might be that this forest soil had the lowest pH among the soils used (Table 1). It is well known that fungi play a more important role than bacteria in organic matter decomposition in acidic forest soils (20). Kuroki and Hayano (8) also reported that the production of phosphodiesterase was primarily attributed to fungi in a forest soil with a pH of 5.5. No effect of Truog-P was observed on whether bacteria or fungi dominated phosphatase production.

While it is generally believed that bacteria are the main source of alkaline phosphatase activity in soils (5), our findings showed that fungi might be the primary contributors in at least some other soils. Although this study included a limited number of samples, these results indicate that phosphatase production in acidic forest soils (including that of alkaline phosphatase) might be primarily mediated by fungi. The results of this study also indicate that, within a given soil, the same microbial group (either bacteria or fungi) tended to dominate the production of all three phosphatase types. The origins of these phosphatases will need to be confirmed in the future by proteomic analyses. Finally, the identity of the dominant producer might be dependent on soil pH.