Decoupling of cellulose decomposition and glucose mineralization in volcanic forest soils

ABSTRACT Decomposition of organic matters in volcanic soils (ando soils) is generally slowed by sorption onto short-range-order minerals, while decomposition of non-charged substrates such as cellulose and glucose is hypothesized to be promoted by high microbial biomass and nitrogen availability, irrespective of short-range-order minerals. To analyze factors regulating decomposition activities of cellulose and glucose, we measured the decomposition rates of cellulose filter papers and the mineralization rates of 14C-labeled glucose in five volcanic soils in Japan. Glucose mineralization activities increased with increasing microbial biomass C and N, while cellulose decomposition activities (standardized with cumulative temperature) were not related to microbial biomass C or N. Cellulose decomposition activities increased with increasing ratio of soil available N relative to microbial biomass N (microbial N availability), while they decreased with decreasing soil pH and with increasing fungal activities. Soils with relatively high soil pH and microbial N availability exhibit the relatively high potentials of cellulose decomposition. Because cellulose decomposition, rather than glucose mineralization, is a rate-limiting step, soil pH and microbial N availability, rather than microbial biomass, could primarily regulate decomposition rates of cellulose and glucose in volcanic soils.


Introduction
Volcanic soils (esp., ando soils) have the largest soil organic carbon (SOC) stocks among major forest soil types in Japan (Imaya et al. 2010;Takahashi et al. 2010). Decomposition of organic matter (OM) in volcanic soils is generally slowed by chemical sorption of OM onto short-range order (SRO) aluminum (Al) and iron (Fe) oxides (e.g., allophane and imogolite) as well as the protection of OM by its occlusion into aggregates (Chevallier et al. 2010;Fujii et al. 2019). On the other hand, the decomposition of non-charged substrates such as cellulose and glucose might differ from that of the charged OM (e.g., organic acids) due to the lack of sorption (Jones and Edwards 1998;Hayakawa et al. 2018). Understanding the regulating factors of microbial activities to decompose a variety of OM is important for assessing the effects of soil types and land use on CO 2 emission in forest soils.
Volcanic soils rich in organic matter tend to have high microbial biomass in forests (Inubushi and Kong 2014). This is supported by the positive correlations between soil carbon (C) and microbial biomass at global and local scales (Inubushi, Sakamoto, and Sawamoto 2005;Xu, Thornton, and Post 2013). The high microbial biomass and metabolites could be the potential sources of SOM in volcanic soils (Miltner et al. 2012;Kallenbach, Frey, and Grandy 2016). On the other hand, the high microbial biomass needs to be supported by supply of fresh and soluble substrates, rather than recalcitrant SOM (Van Hees et al. 2005), because the majority of SOM is highly resistant to decomposition in volcanic soils due to sorption of organic matters onto SRO minerals (Shirato, Hakamata, and Taniyama 2004;Fujii et al. 2019). This raised a question whether the high microbial biomass can be supported by the supply of fresh and soluble substrates in volcanic soils.
Root exudates and litter-soluble constituents (e.g., sugars) can be directly absorbed by microbes, but the majority of organic matter is present in insoluble polymers (Van Hees et al. 2005). Microbial biomass requires the supply of soluble monomers through the solubilization of polymers (Allison, Wallenstein, and Bradford 2010). For example, cellulose must be depolymerized to glucose by a set of cellulases, including endocellulase, exocellulase, and cellobiose dehydrogenase (Linkins et al. 1990). Although both cellulose decomposition and glucose mineralization are driven by microbial community, decomposition rates could vary depending on climate and soil properties such as microbial biomass and composition (fungi or bacteria), soil pH (optimal pH range for enzymes), and soil available N (substrates for enzyme production) (Hayakawa et al. 2011). It remains unclear whether the high microbial biomass leads to the high activities of cellulose solubilization and glucose mineralization in volcanic soils.
Cellulose decomposition rates measured in non-volcanic forest soils in our previous study (Hayakawa et al. 2014) were much lower than those reported in the adjacent agricultural soils with higher pH and N availability (Fujii, Inagaki, and Hayakawa 2020). However, unlike non-volcanic forest soils, volcanic forest soils rich in organic matter and SRO minerals might exhibit the high activities of cellulose solubilization, because the high acid neutralizing capacity of SRO minerals results in higher soil pH and N availability that are favorable for high activity and production of cellulases Urakawa et al. 2016;Averill and Waring 2018). The ratelimiting steps and factors of two decomposition steps (polymer solubilization and monomer mineralization) might differ between volcanic and non-volcanic soils .
We used cellulose and glucose as standard substrates of polymers and monomers. While plant litter comprises a heterogeneous mixture of polymers, cellulose is a major polymer in plant materials (Berg and McClaugherty 2003). Cellulose paper is useful for comparing soil microbial responses to polymers without the variation caused by heterogeneity of substrate (Drewnik 2006;Kim, An, and Byun 2010;Hayakawa et al. 2014Hayakawa et al. , 2020. Glucose in soil solution is one of the main substrates for heterotrophic respiration (Fujii et al., 2010) and 14 C labeled glucose is a useful monomer substrate to compare microbial activities between soils due to negligible sorption onto SRO minerals (Jones and Edwards 1998). Due to the low detection limit of 14 C, 14 C tracer experiments could be conducted at the lower substrate availability that is closer to the field condition, compared to 13 C tracer experiment. We compared the decomposition activities of cellulose and glucose in soils that differ in microbial biomass, pH, and N availability to test (1) whether volcanic soils with higher pH and N availability exhibit the higher activities of cellulose solubilization and (2) whether rate-limiting factors differ between two-step decomposition processes (cellulose solubilization and glucose mineralization) in volcanic soils.

Site description
To test whether volcanic soils with higher microbial biomass, pH, and N availability exhibit the higher activities of cellulose solubilization, the soils with different microbial biomass, pH, and N availability were selected from four temperate forest sites and one reference cropland site in Japan (Table 1). Cropland site was included to have the wide variation in soil pH and N availability, as discussed later (Table 2). Two of the forest sites, in Appi (AP) and Morioka (MR) in Iwate Prefecture, were dominated by beech (Fagus crenata). The other two forest sites, located in Katsura (KTR) and Tsukuba (TKB) in Ibaraki Prefecture, were dominated by Japanese cedar (Cryptomeria japonica) plantations. The cropland site was located in a green onion field in Ushiku (USK), Ibaraki Prefecture. All soils were influenced by volcanic ash deposition but are classified as Andisols (MR, and USK) or Inceptisols (AP, KTR, and TKB) based on the abundance of oxalate-extractable Al and Fe minerals (Al o +1/2Fe o ) ( Table 2; Soil Survey Staff 2014). Inceptisol soils were influenced by other parent materials (i.e., sedimentary rocks in AP and KTR and biotite gneiss in TKB). All sites are characterized by a temperate humid climate, and the mean annual air temperatures and annual precipitation are 6.1-14.3°C and 1207-1449 mm yr −1 , respectively. Apart from high N fertilization at cropland USK (200 kg N ha −1 yr −1 ), TKB receives greater atmospheric N deposition (20 kg N ha −1 yr −1 ) compared to the other forest sites (<10 kg N ha −1 yr −1 ) (Inagaki et al. 2015).

Soil sampling
Soil samples (A horizon, 0-5 cm) were collected from five pits at each site. The distance between each pit was 10 m. Samples were stored at 4°C in plastic bags prior to analysis and sieved (<2 mm) to eliminate litter, roots, and stones. A subsample of field-moist soil was used to measure biological properties in a week, while a second, air-dried subsample was used to measure physicochemical properties.

Soil physicochemical properties analyses
The physicochemical properties of air-dried soil samples were assessed using the following methods. Soil pH was measured with glass electrode (Horiba, 9615S-10D) using a soil-tosolution (H 2 O or 1 M KCl) ratio of 1:5 after shaking for 1 h. Total C and nitrogen (N) concentrations were determined using an NC analyzer (NC-22F SUMIGRAPH; Sumika Chemical Analysis Services, Osaka, Japan). Clay content (<2 µm) was estimated using the pipette method after H 2 O 2 pre-treatment for organic matter removal and dispersion (Gee and Bouder 1986 1966). The concentrations of Fe, Al, and Si in soil extracts were determined using inductively coupled plasma atomic emission spectrometry (SPS1500; Seiko Instruments, Chiba, Japan). Microbial biomass of C and N (MBC and MBN, respectively) was determined using the chloroform fumigation-extraction method (Vance, Brookes, and Jenkinson 1987). The soluble C and N of the fumigated and non-fumigated soil samples were extracted with 0.5 M K 2 SO 4 (soil to solution ratio of 1:5) and were determined using a total organic carbon and nitrogen analyzer (TOC-V CSH and TNM-L; Shimadzu, Japan). A kEC factor of 0.45 (Wu et al. 1990) and a kEN factor of 0.54 (Jenkinson 1988) were used to calculate MBC and MBN, respectively. Basal soil respiration rates [steady-state respiration rate in soil derived from native organic matter (Pell, Stenstrom, and Granhall 2005)] were determined in the laboratory by measuring carbon dioxide (CO 2 ) emission from field-moist soils (equivalent to 10 g dry weight soil) incubated in the dark at 25°C for 3 h in 100-mL Erlenmeyer flasks sealed with silicone rubber septa. The experiments were carried out in five replicates. Evolved CO 2 was collected in glass vials using a syringe, measured with an infrared CO 2 controller (ZFP9, Fuji Electric Instruments Co., Saitama, Japan), and expressed as μg C kg −1 h −1 . An indicator of soil C decomposability was obtained by dividing the amount of basal respiration C by total soil C.

Measurement of cellulose decomposition rates under field conditions
The decomposition rates of cellulose were estimated from the mass loss of cellulose filter papers buried in the soil (Hayakawa et al. 2014). Single pieces of cellulose filter paper (Advantec No.6, 55-mm diameter) were enclosed in 65 × 65-mm nylon mesh bags (mesh size = 100 μm) to exclude insects and worms, then buried into the surface mineral soil (A horizon, 5 cm depth) to minimize disturbance to the soil structure. The filter paper samples (5 sampling intervals × 5 replicates) were collected at each site once or twice a month. The remaining filter papers in the mesh bags were dried at 70°C for 24 h and weighed following careful removal of soil particles. The remaining weight of the substrate was calculated on an ash-free basis by subtracting the weight of the soil adhering to the substrate, which was estimated by dry combustion (600°C, 4 h). The proportion of the remaining substrate was calculated by dividing the final substrate weight by the initial weight. The data were fitted to the first-order kinetics: where R r is the remaining proportion of the substrate (%), R i is the initial proportion of the substrate (i.e., 100%), k time is a constant representing decomposition rates (yr −1 ), and t is time (yr). To normalize temperature effects, decomposition rate constants were also calculated based on cumulative temperature: where k cumulative temp. is the decomposition rate constant (degree day −1 ) and t is the cumulative soil temperature (degree day). Cumulative soil temperature (degree day) was calculated by summing the average daily soil temperature above 0°C. Soil temperature at 5 cm depth were measured using temperature loggers (Thermochron, SL type).

Glucose mineralization rates and fungal respiratory activities
We added 100 µL of 14 C-radiolabeled glucose solution (U-14 C, 0.4 GBq mmol −1 , specific activity = 0.17 kBq mL −1 , pH 4.5; American Radiolabeled Chemicals, St. Louis, MO, USA) to 1 ± 0.02 g of field-moist soil in 50-mL polypropylene vials. The initial solution concentrations of glucose were 50 µM, which was determined based on the actual substrate availability in soil solution (Fujii et al., 2018). Glucose addition amounted to 0.1-0.5% of soil C. Following the addition of glucose, the soil was gently shaken to ensure mixing and then incubated at 25°C in sealed vials. The 14 C-CO 2 produced by the biodegradation of the added substrate was collected in a plastic scintillation vial containing 1.0 mL of 1 M NaOH placed on top of the soil and separated by a spacer. The 14 C-CO 2 content trapped in NaOH was determined by liquid scintillation counting (LSC-3050, Aloka Co., Tokyo, Japan) using alkalicompatible scintillation fluid (Hionic-Fluor; Perkin Elmer, Waltham, MA, USA). The production of 14 C-CO 2 was measured during the initial linear phase of decomposition (1 h), which was confirmed by pilot experiment. The experiments were conducted in five replicates.
To obtain rough estimates of soil fungal activity, we measured the relative contribution of fungi to glucose-induced respiration in soil using the selective inhibition method Domsch 1973, 1978;Joergensen and Wichern 2008). Samples (10 g) of field-moist soil were placed in vials and amended with a solution of 14 C-labeled glucose (5 mg glucose g −1 soil; U-14 C, 0.4 GBq mmol −1 ; American Radiolabeled Chemicals) with or without cycloheximide (a fungal respiratory inhibitor, equivalent to 24-48 mg g −1 soil) and incubated at 22°C for 24 h (Fujii et al. 2012). We collected and determined the 14 C-CO 2 in a 1 M NaOH solution using liquid scintillation counting. The contribution of fungi to total microbial respiration (%) was calculated using the differences in 14 CO 2 evolution rates of the glucose-amended soils with and without the inhibitor. The experiments were carried out in five replicates.

Statistical analyses
All data are expressed as the mean ± standard error (SE) and the combined SE of five replicates (Taylar 1997;Zar 1999). The cellulose decomposition data were fitted to a single exponential decay function using the least-squares technique in SigmaPlot 14.5 (SYSTAT Software Inc., Point Richmond, CA, USA). We tested for statistical differences among mean values of MBC, MBN, available N, fungal respiratory activity, basal respiration, and decomposition rates between sites using an analysis of variance and a significance level of P < 0.05. A Pearson's correlation coefficient test was used to examine relationships between the rate constants and soil properties. The differences in linear regression slopes of the rate constants related to soil pH were tested using analysis of covariance (ANCOVA) between volcanic soils and non-volcanic soils. Analyses were performed using SigmaPlot 14.5.

Physicochemical and biological properties of soils
Soil C concentrations were not correlated with clay contents nor with oxalate-extractable Al and Fe (abundance of SRO minerals) ( Table 2). Soil pH was higher in the USK cropland soil than in the forest soils, while soil C concentrations and microbial biomass were lower (Tables 2 and 3). Microbial biomass C corresponded to 0.2-0.8% of soil C concentrations. Basal respiration rates were higher in the USK cropland soil and the KTR and TKB cedar forest soils than in the AP and MR beech forest soils (Table 3). The proportions of basal respiration C relative to bulk soil C in the forest soils (0.8-3.2 × 10 −3 % h −1 ) were lower than in the USK cropland soil (6.7 × 10 −3 % h −1 ; Table 3). Soil available N (K 2 SO 4 extractable) was the higher in the TKB cedar forest site than in the other sites (Table 3). Soil available N per microbial biomass N followed the order: USK (cropland) > TKB (cedar) > KTR (cedar) > MR (beech) > AP (beech) ( Table 3).

Rate constants of cellulose decomposition in soils
Within the initial 200 days of incubation, the weight of the cellulose filter papers decreased by 87-100% at the MR, KTR, TKB, and USK sites, but only by 50% at the AP beech forest site (Figure 1). When the remaining proportion of cellulose over time was fitted to Eq. 1 (Figure 1; R 2 >0.88, P < 0.01), the rate constants (k time ) varied widely, from 1.0 to 19.4 yr −1 ( Table 3). The rate constants decreased in the following order: USK (cropland) > TKB (cedar) > KTR (cedar) > MR (beech) > AP (beech) (Figure 1; Table 3). The rate constants (k time ) were positively correlated with soil temperatures during the field incubation study (Figure 2). The rate constants (k cumulative temp. ) of cellulose decomposition differed among sites even after conversion of the x-axis from time (day) to cumulative soil temperature (degree day) (Figure 3). When the data were fitted to Eq. 2 (Figure 3;R 2 >0.90, P < 0.01), the rate constants (k cumulative temp. ) varied from 4.1 to 19.9 (×10 −4 degree day −1 ) ( Table 3). The rate constants decreased in the following order: USK (cropland) > TKB (cedar) > MR (beech) > KTR (cedar) > AP (beech) (Figure 3; Table 3). The single correlation analyses between the rate constants (k cumulative temp. ) and soil properties showed that k cumulative temp. values and microbial biomass C are not significantly correlated ( Figure 4a); however, the rate constants increased with the proportion of basal soil respiration relative to soil C increased (Figure 4b). No correlations were found between the rate constants (k cumulative temp. ) and the other soil properties (soil C and N, C/N, soil available N, soil particle size percentages, oxalate extractable Al and Fe, microbial biomass, fungal respiratory activity, and basal respiration), except for the ratio of soil available N per MBN that correlated positively with the rate constants (k cumulative temp. ) (Figure 4c). Using our data and those from the published data (Hayakawa et al. 2014) that comparable data were obtained based on the same method (Table S1), the rate constants (k cumulative temp. ) were positively correlated with soil pH (Figure 4d). An ANCOVA for the linear regression slopes of the rate constants related to soil pH with factor of soil types (volcanic soils vs. non-volcanic soils) exhibited no significant (P > 0.05) effects of soil types nor significant (P > 0.05) interaction effect of soil pH × soil types. Thus, soil pH was extracted as a single parameter regulating the rate constants in our study (Figure 4d).

Glucose mineralization rates and fungal respiration
Between 4.8% and 7.9% of the added 14 C-labeled glucose decomposed within 1 h of addition (Table 3). Glucose mineralization rates were significantly higher in the forest soils (P < 0.05) than in the USK cropland soil (Table 3). Glucose mineralization rates were positively correlated with microbial biomass C (Figure 5a) or microbial biomass N (not shown). Glucose mineralization rates have no significant relationship with the rate constants (k cumulative temp. ) of cellulose decomposition (Figure 5b) or the other soil properties [soil C and N, C/N, soil available N, soil particle size percentages, oxalate extractable Al and Fe (Figure 5c), microbial biomass, fungal respiratory activity, and basal respiration]. Glucose mineralization rates in soils with and without cycloheximide (fungal respiratory inhibitor) indicated that the relative contribution of fungi to glucose-induced respiration (fungal respiratory activity) ranged from 16% to 36% in our study (Table 3). Using our data and those from the published sources (Fujii et al. 2013;Hayakawa et al. 2014; Table S1), the fungal respiratory activity increased with decreasing soil pH (Figure 6a). The rate constants (k cumulative temp. ) of cellulose decomposition were negatively correlated with fungal respiratory activity (Figure 6b).

pH regulation of cellulose degradation in volcanic soils
We tested whether the high microbial biomass leads to the high activities of cellulose solubilization or not. Cellulose test has advantages in analyzing environmental controls on the decomposition of standard substrate by canceling the effects of substrate quality between sites (Drewnik 2006). Cellulose decomposition rates under field conditions are primarily regulated by soil temperatures (Figure 2), consistent with Drewnik (2006) and Hayakawa et al. (2014). We standardized the effects of soil temperatures on decomposition rates by conversion of Results are the mean ± standard error of five replicates. Within each column, values denoted by the same letter do not differ significantly (P > 0.05). a K 2 SO 4 extractable N from the non-fumigated soil. Soil available N per microbial biomass N (MBN) was calculated by dividing available N by MBN. b 100% -(the proportion of 14 C-labelled glucose-induced respiration in soil with cycloheximide (fungal respiratory inhibitor) relative to respiration rate in soil without cycloheximide) c Basal respiration was measured for the incubation (25ºC) of the field-moist soils in laboratory. d Respired C/ Soil C denotes the proportion of basal respiration rate relative to the bulk soil C. e Comparison of decomposition rates of cellulose under field condition and those of 14 C-labelled glucose under laboratory incubation.
f Decomposition rate constants of cellulose obtained from fitting to the first order kinetics based on time or cumulative temperature. the x-axis from time (day) to cumulative soil temperature (degree day) (Figure 3; Curtin and Fraser 2003;Hayakawa et al. 2014). However, the wide variation in k cumulative temp. values suggests that rates of cellulose decomposition are regulated by the other environmental factors as well as soil temperature ( Table 2). The decreased k cumulative temp values at lower soil pH (Figure 4d) are consistent with previous reports that cellulase activity sharply declined as pH decreased from 5.5 to 4.0 (Deng and Tabatabai 1994;Criquet 2002) partly due to deactivation of enzymes by Al 3+ (Miltner and Zech 1998;Scheel et al. 2008) and that microbial activity generally decreases at low pH due to increased Al 3+ toxicity or high H+ concentrations (Kemmitt et al. 2006). High microbial biomass does not always lead to high microbial activities of cellulose decomposition (Figure 4a). This contrasts with glucose mineralization activity, which increases with increasing microbial biomass (Figure 5a). Low soil pH generally limits bacterial activity, and microbial communities could be dominated by fungi in acidic forest soils (Bååth and Anderson 2003;Rousk, Brookes, and Baath 2009). This is consistent with the negative correlation between fungal respiratory activities and pH at our sites (Figure 6a). Fungi are known to be the major producers of cellulases and degrade cellulose more rapidly than bacteria under pure culture incubation (Lynd et al. 2002), but cellulose decomposition rates under the field condition decrease with increasing fungal dominance in our study (Figure 6b), likely due to decreased cellulase activity at lower soil pH (Criquet 2002). The relatively high soil pH, rather than high microbial biomass, is an important factor for high activities of cellulose degradation (Figure 4d). Soil pH tends to be acidic under humid forest environments (Slessarev et al. 2016), but the high acid buffering capacities derived from the high amounts of SRO minerals could mitigate acidification of the volcanic soils, compared to the non-volcanic soils (Table 2; Figure 4d; Funakawa, Hirooka, and Yonebayashi 2008). The relatively high pH conditions are favorable for high activities of cellulose decomposition in the volcanic soils (Figure 4d).

Different patterns of cellulose and glucose decomposition activities in volcanic soils
Substrates are supplied as both polymers and monomers, which differ in that polymers require solubilization by enzymes prior to the uptake of monomers by microbes (Allison, Wallenstein, and Bradford 2010;Hayakawa et al. 2014). We observed substantial differences in decomposition activities between polymer (cellulose) and monomer (glucose) and  between soils, respectively (Table 3; Figure 5b). When the rates were transformed into the same unit for rough comparison, the decomposition rates of cellulose (0.01% to 0.06% h −1 ) are far slower than those of glucose (4.8% to 7.9% h −1 ) ( Table 3). This is consistent with the fact that the rate-limiting step of organic matter is depolymerization to monomers, rather than microbial  . When microbes are limited by energy rather than substrate (polymer) and N for enzyme production, the microbial community is assumed to produce cellulolytic enzymes to obtain glucose (Allison and Vitousek 2005). As also seen in Fujii, Inagaki, and Hayakawa (2020), the higher soil pH and N fertilization, as well as high temperatures, could explain cellulose decomposition activities observed in the cropland soil, despite the lower microbial biomass (Figure 4c). The higher N deposition in the TKB forest soil results in higher soil N availability/MBN and the higher cellulose decomposition activities, compared to the other three forest sites (Table 3). Although N availability is not a determining factor such as pH (Figure 4d), the higher N availability might be favorable for production of cellulases to decompose cellulose even under lower pH of TKB soil (Allison and Vitousek 2005; Table 3). On the other hand, microbial activities of cellulose decomposition might be suppressed by lower pH or microbial N availability in the forest soils (AP, MR, KTR; Figure 4c). Because cellulose decomposition, rather than glucose mineralization, is a ratelimiting step (Table 3), soil pH and microbial N availability, rather than microbial biomass or composition (fungal activity), could primarily regulate decomposition rates of cellulose and glucose in volcanic soils ( Figures. 4c-d; Figure 6b).

Implication for carbon cycles in volcanic soils
The high microbial biomass or necromass could be the potential source or precursor of SOM (Miltner et al. 2012;Kallenbach, Frey, and Grandy 2016), but the high microbial biomass needs to be supported by the supply of fresh and labile substrates. Although the data of cellulosic-C concentrations are not available for the studied soils, but the existing literature showed that the cellulosic-C accounts for low proportions of soil C [1.5% to 13.2% from Ishizuka et al. (2006) and Hayakawa et al. (2014)] and that the majority of SOM is highly resistant to decomposition in volcanic soils (Shirato, Hakamata, and Taniyama 2004). The high stability of SOM in forest soils is also supported by our observation of the lower proportions of basal respiration/soil C, compared to the cropland soil (Table 3). We have previously reported using the same soil samples that mineralization rates of the charged substrates (e.g., organic acids) could be slowed by sorption onto SRO minerals in the volcanic soils (Fujii et al. 2019).
Regarding non-charged substrates, cellulose solubilization and glucose mineralization rates are independent from the abundance of SRO minerals (Figure 5c). Rather, SRO minerals in volcanic soils contribute to mitigating acidification of volcanic soils (Funakawa, Hirooka, and Yonebayashi 2008) and enhance cellulose decomposition activities (Figure 4d). The glucose supply to microbial biomass in volcanic soils could not be limited by sorption on to SRO minerals, but the supply rates via cellulose solubilization are highly variable among volcanic soils depending on pH and microbial N availability (Figure 4c-d), in addition to the small cellulose pool size of cellulosic-C (Hayakawa et al. 2014) and physical occlusion within aggregates (Chevallier et al. 2010). The discrepancy between the low activity of cellulose decomposition and the high microbial biomass in the acidic forest soils (e.g., AP; Table 3) suggests that high microbial biomass needs to be supported by the other C sources (e.g., root exudates, lignin, and non-cellulosic SOM) as well as cellulose.

Conclusions
It is well known that organic matter mineralization is limited by sorption onto SRO minerals in volcanic soils, but we found that decomposition of non-charged substrates such as cellulose and glucose are independent from the abundance of SRO minerals. Glucose mineralization activities increased with increasing microbial biomass, while cellulose decomposition activities (standardized with cumulative temperature) were not related to microbial biomass. Cellulose decomposition activities increased with increasing ratio of soil available N relative to microbial biomass N (microbial N availability), while they decreased with decreasing soil pH. The soils with relatively high soil pH and microbial N availability exhibit the relatively high potentials of cellulose decomposition. Because cellulose Present study Reference data Figure 6. Relationships (a) between soil pH and fungal respiratory activity and (b) between fungal respiratory activity and rate constants of cellulose decomposition. The reference data was Hayakawa et al. (2014). Bars indicate standard errors (N = 5).
decomposition, rather than glucose mineralization, is a ratelimiting step, soil pH and microbial N availability, rather than microbial biomass, could primarily regulate decomposition rates of cellulose and glucose in volcanic soils.