Introduction

Soil enzymes, acting as biological catalysts, can accelerate chemical reactions involving soil organic matter (Jin et al. 2009). Because soil enzymes are main regulators of the biochemical processes within the soil environment, they play a critical role in the development of soil characteristics. Soil characteristics are determined by the interplay between enzymes and substances produced by a wide range of animals, plants, microorganisms, and their residues (Zhang et al. 2005). Soil enzymes, which are intimately associated with decomposition of organic matter within ecosystems, nutrient cycling, and energy transfer, greatly impact resulting environmental quality (Johansson et al. 2000; Böhme et al. 2005; Jin et al. 2009; Kaiser et al. 2010; Burke et al. 2012). Therefore, soil enzyme activities can be used as an evaluation index of microbial activity and soil fertility (Monreal and Bergstrom 2000; Alvarez and Guerrero 2000). Changes in enzymes are known to alter the availability of nutrients for uptake by plants, and these changes are potentially sensitive indicators of soil quality (Albiach et al. 2000; Aon et al. 2001; Aon and Colaneri 2001; Sinsabaugh et al. 2009). Therefore, there is practical significance to research soil enzyme activities, especially the impact of rhizosphere soil enzymes on the mechanisms that drive ecosystem degradation and restoration (Nausch and Nausch 2000; Fioretto et al. 2001).

The rhizosphere is a special ecological zone that is impacted by interactions between plants, soil, and microorganisms (Orwin et al. 2010; De Graaff et al. 2010; Bell et al. 2014). Currently, a highly active area of biochemical research, the rhizosphere is an important site of material cycling and energy flow within the soil. Because soil properties depend on secretion of various organic substances and absorption of unbalanced nutrients by tree roots, changes in soil properties within the rhizosphere are early indicators of changes in tree roots. Because the physiological activities of tree roots are determined by their own genetic characteristics and because activities vary greatly between the different developmental stages of trees, the properties of rhizosphere soil are affected directly by changes in tree root characteristics as a forest changes with time (Kardol, et al. 2010; Sanaullah et al. 2011).

Nothotsuga longibracteata (W.C. Cheng) Hu ex C.N. Page is a relict tree species endemic to China, having originated in the Tertiary Period, belonging to the genus Nothotsuga in the family Pinaceae. It is mainly distributed within subtropical mountain areas in southern China, including Fujian, Jiangxi, Hunan, Guizhou, Guangdong, and Guangxi Provinces. Nanling Mountain and the Daiyun Mountains are the main distribution areas of N. longibracteata, and trees mainly exhibit a patchy distribution tendency (Qiu et al. 2013). This species plays a crucial part in the promotion of forest succession, maintenance of balance and stability of ecosystems, and water conservation (You et al. 2013b, c). The distribution area of N. longibracteata forest is 186.7 hm2 in the Tianbaoyan National Nature Reserve. The primary area of pure N. longibracteata forest is 20 hm2, and this forest community is the largest area of N. longibracteata pure forest in China (Xiao et al. 2012). The community is composed of large areas of primeval forest and has vital significance for researching ecosystem responses to global climate changes (You et al. 2013a).

The current study investigated protease, invertase, cellulase, urease, and acid phosphatase activities of rhizosphere and bulk soils from six different N. longibracteata forest communities within the Tianbaoyan National Nature Reserve. The forest types included five mixed forest types (N. longibracteata + Phyllostachys pubescens forest, N. longibracteata + Schima superba forest, N. longibracteata + Rhododendron simiarum forest, N. longibracteata + Cunninghamia lanceolata forest, and N. longibracteata + Cyclobalanopsis glauca forest) and one pure forest (N. longibracteata pure forest).

Materials and methods

Site description for sample collection

The study site is located in the Tianbaoyan National Nature Reserve (25°50′51″–26°01′20″N, 117°28′3″–117°35′28″E), Sanming City, Fujian Province of China (Fig. 1), and covers a total area of 11,015 hm2. The altitudinal gradient of Tianbaoyan Reserve is from 580 to 1604 m and is within the Daiyun Mountain Range. The Tianbaoyan National Nature Reserve has a mid-subtropical oceanic monsoon climate with a mean annual temperature of 15 °C and a frost-free period of about 290 days. The extreme high temperature is 40 °C and extreme low is −11 °C. Annual average precipitation is 2039 mm, occurring mostly between May and September. Mean annual relative humidity is greater than 80%. The reserve has four distinct seasons with a warm and humid climate, where the light, heat, and water conditions are sufficient for forests growth. Devonian and Jurassic sedimentary rocks and granite are exposed in the reserve. Upland red, zonal soils occur below 800 m asl. Upland yellowish soils occur between 800 and 1350 m asl and upland yellow soils occur above 1350 m asl. There is abundant species diversity in the reserve, retaining a large number of natural R. simiarum, N. longibracteata, and Cryptomeria japonica forests. All the three types of forest communities are of high conservation value.

Fig. 1
figure 1

Geographic location of the Tianbaoyan National Nature Reserve

Full size image

Procedure for soil sampling

All soil samples were collected from the Tianbaoyan National Nature Reserve in July 2012 from a depth of 0–20 cm within the topsoil layer. Samples for this study were collected from six N. longibracteata forests within Tianbaoyan National Nature Reserve (N. longibracteata + P. pubescens, N. longibracteata + S. superba, N. longibracteata + R. simiarum, N. longibracteata + C. lanceolata, N. longibracteata + C. glauca, and pure N. longibracteata forest). Six similar stands were chosen from the forests as study sites; each stand contained three plots per stand with each plot defined by dimensions 20 × 30 m. Longitude and latitude, altitude, slope, slope aspect, and crown coverage in each plot were recorded. The site characteristics of the forest types studied are listed in Table 1. Based on the measurement of all individual trees within the population of each plot, three average trees were chosen as standard trees for rhizosphere soil sampling. Soil adhering to fine roots <2 mm in diameter was defined as rhizosphere soil (Garcia et al. 2005). The remaining soil was defined as bulk soil. All soil samples were collected from a depth of 0–20 cm of the topsoil layer. The rhizosphere soil of N. longibracteata was separated from bulk soil by shaking the excavated fine root fragments gently. In each plot, five randomly located soil cores were taken from soil near three average trees. The soil samples were immediately mixed into a single sample for each plot. Root fragments remaining in the rhizosphere and bulk were carefully removed with sterilized forceps. Samples were then air-dried for 10 days at room temperature until they achieved constant mass. Samples were then used for soil enzyme assays.

Table 1 Site characteristics of the studied forest communities
Full size table

Enzyme assays

Protease activity was measured according to Zhang et al. (2005). Briefly, 4 g of air-dried soil was incubated for 24 h at 30 °C with 20 mL of 1% (w/v) casein and 1 mL of toluene, then 2 mL of 0.1 mol L−1 sulfuric acid and 12 mL of 20% (w/v) sodium sulfate were added and the suspension centrifuged for 15 min (6000 rpm). The amino nitrogen released was reacted with 1 mL 2% (w/v) ninhydrin, and the soil suspension was heated in a boiling-water bath for 10 min, and the absorbance read at 500 nm.

Invertase activity was measured by incubating 5 g of air-dried soil for 24 h at 37 °C with 15 mL of 8% (w/v) sucrose, 5 mL of phosphate buffer at pH 5.5 and 5 drops of toluene. The glucose released was reacted with 3,5-dinitrosalicylic acid, and absorbance measured at 508 nm (Zhang et al. 2005).

Cellulase activity was measured by adding 1 mL of phosphate buffer at pH 5.5 and 0.3 mL of toluene to 5 g of each soil sample. Then 3 mL of 3,5-dinitrosalicylic acid was added, and the soil suspension was heated in a boiling water bath for 5 min. Next, each soil sample was cooled under running water for 3 min. The glucose released was reacted with the 3,5-dinitrosalicylic acid, then measured at 540 nm (Zhang et al. 2011).

Urease activity was assayed by incubating 5 g of air-dried soil with 1 mL of toluene, 10 mL of 10% (w/v) urea solution, and 20 mL of citrate buffer at pH 6.7 for 24 h at 37 °C. Amino nitrogen was measured at 578 nm, and results were expressed as NH3 released in mg NH3-N per 1 g of soil at 24 h (Zhang et al. 2011).

Acid phosphatase activity was determined as described by Zhang et al. (2005). Air-dried soil (5 g) was incubated for 24 h at 37 °C with 2.5 mL toluene and 20 mL 0.5% (w/v) disodium benzene phosphate dissolved in acetate buffer at pH 5. The phenol released was measured at 510 nm and results expressed as milligrams phenol per 1 g of soil at 24 h. All absorbances were measured using a UV20600 Unicam UV–Vis spectrometer (Unicam, USA).

Statistical analyses

All reported results were averages of three replicates. Analysis of variance was used to detect whether the rhizosphere soil varied significantly relative to bulk soil. The least significant difference was used to compare means of the different enzyme activities measured (p < 0.05 and p < 0.01). A principal component analysis (PCA) was used to analyze most of the variance in the data after reducing the number of variables to a few uncorrelated components. PCA was used to identify groups of soil chemical properties and correlate them to enzyme activities in this study. For the PCA, the data was standardized to zero mean and unit variance, and analysis was conducted using the correlation matrix. Statistical procedures were carried out using the SPSS 21.0 software package for Windows (IBM, USA). The rhizosphere effect is usually described using the ratio of values from the rhizosphere and non-rhizosphere zones, namely the rhizosphere effect value (R/S). PCA ordination was implemented using CANOCO software version 5 for Windows (Wageningen UR, Netherlands).

Results

Enzyme activities

Figure 2a–e shows protease, invertase, cellulase, urease, and acid phosphatase activities of both rhizosphere and bulk soils of the six N. longibracteata forest communities. It is clear that the rhizosphere had higher enzyme activities than bulk soils. The highest protease activities were found in both rhizosphere and bulk soils of N. longibracteata + R. simiarum forest (96.047 μg g−1 and 87.230 μg g−1, respectively), and the lowest in bulk soils of N. longibracteata + C. lanceolata forest (92.613 μg g−1) and rhizosphere of N. longibracteata + P. pubescens forest (85.523 μg g−1) (Fig. 2a). Rhizosphere invertase activities varied from 16.927 mg g−1 in N. longibracteata + S. superba forest to 15.706 mg g−1 in P. pubescens forest; bulk soil invertase activities varied from 13.313 mg g−1 in N. longibracteata + C. glauca forest to 12.479 mg g−1 for N. longibracteata + S. superba forest (Fig. 2b). Cellulase activity tended to be the highest in the rhizosphere soil in the pure N. longibracteata forest; the lowest activity was observed in N. longibracteata + C. lanceolata forest. In bulk soil, the highest cellulase activity was observed in N. longibracteata + S. superba forest (0.861 mg g−1) and the lowest in N. longibracteata + R. simiarum forest (0.744 mg g−1) (Fig. 2c). Urease levels ranged from 0.794 mg g−1 in the rhizosphere soil of N. longibracteata + S. superba forest, to 0.677 mg g−1 in pure N. longibracteata forest, while these activities varied from 0.645 mg g−1 in bulk soils of N. longibracteata + C. glauca forest vs. 0.564 mg g−1 for N. longibracteata + R. simiarum forest (Fig. 2d, e shows that the maximum acid phosphatase activity was observed in the rhizosphere soil of N. longibracteata + R. simiarum forest (0.637 mg g−1) and the minimum was observed in soil of pure N. longibracteata forest (0.504 mg g−1), while in bulk soil the maximum was observed for N. longibracteata + S. superba forest (0.324 mg g−1) and the minimum observed for N. longibracteata + R. simiarum forest (0.288 mg g−1). There was a significant difference (p < 0.05) between the rhizosphere and bulk soils within the same forest. The rhizosphere effect values of all enzyme activities in all six forests were all >1, indicating that rhizosphere effects were obvious.

Fig. 2
figure 2

Comparison of protease (a), invertase (b), cellulase (c), urease (d), acid phosphatase (e) activities (mean ± SE, n = 6) in rhizosphere and bulk soils of different N. longibracteata forest types. Notes: N.l + P.p: N. longibracteata + P. pubescens forest; N.l + S.s: N. longibracteata + S. superba forest; N.l + R.s: N. longibracteata + R. simiarum forest; N.l + C.l: N. longibracteata + C. lanceolata forest; N.l + C.g: N. longibracteata + C. glauca forest; N.l: pure N. longibracteata forest; different letters denote significant differences between rhizosphere and bulk soil in the same forest (p < 0.05)

Full size image

Relationships among enzyme activities in rhizosphere and bulk soils

The canonical correlation analysis elucidated the relationships among rhizosphere soil enzyme activities of different N. longibracteata forests within the Tianbaoyan National Nature Reserve (Table 2). There were significant linear correlations between acid phosphatase and protease activities (p < 0.01) and acid phosphatase and cellulase activities (p < 0.05). However, no significant relationships were found among other rhizosphere soil enzyme activities.

Table 2 Canonical correlation analysis among enzyme activities in rhizosphere soil of N. longibracteata
Full size table

Table 3 shows the canonical correlation analysis among enzyme activities in bulk soils of the six N. longibracteata forests. There were significant linear correlations among bulk soil enzyme activities between protease and acid phosphatase activities (p < 0.05) and between invertase and cellulase activities (p < 0.01). Significant relationships were not found among other enzyme activities.

Table 3 Canonical correlation analysis among enzyme activities in bulk soil of N. longibracteata
Full size table

Principal component analysis

A partial form of PCA was used to estimate the importance of properties variables in the grouping of the studied fields. The data set of rhizosphere and bulk soil chemical properties and enzyme activities in different N. longibracteata forests was analyzed using the PCA. The rhizosphere and bulk soil chemical properties (pH, organic matter [OM], total N [TN], hydrolysable nitrogen [HN], total P [TP], available P [AP], total K [TK], and available K [AK]) were measured in 2014. Table 4 and Fig. 3 show that the first principal component (PC1) accounted for 61.233% of the total variance, and the properties with higher loading included protease, invertase, acid phosphatase, total N, and cellulase activities. The second principal component (PC2) accounted for 15.422% of the total variance and were positively related to total K and available K with negative loadings. The highly loaded properties under the third principal component (PC3) included available P. Other soil chemical properties and enzyme activities contributed relatively low loadings to the PCs (e.g. pH, organic matter, hydrolysable N, total P, and urease) indicating that their variability was much more muted and contributed less to the overall soil chemical properties and enzyme activities variability in the study area.

Table 4 Principal component analysis of rhizosphere and bulk soil chemical properties and enzyme activities of different N. longibracteata forest types
Full size table
Fig. 3
figure 3

Correlation of soil chemical properties and enzyme activities with PCA axes

Full size image

Discussion

Soil enzymes are secreted by the residues of plant roots, soil animals, and their remains and microorganisms. They undoubtedly perform functions critical to the dynamics of nutrient transformation (Zhang et al. 2011). They are considered to be sensitive indicators of soil quality in both natural and agricultural ecosystems due to their rapid response to changes in soil quality (Puglisi et al. 2006). In the past few decades, enzyme activity has been used to evaluate environmental quality indicators such as productivity, fertility, pollution effects, and nutrient cycling potential (Imamura et al. 2006). In general, the soil parameters that were measured in our study (protease, invertase, cellulase, urease, and acid phosphatase) exhibited higher values in the rhizosphere soils than bulk soils, presumably because the rhizosphere induces the synthesis of these enzymes. Levels of soil enzymes have a close relationship between the organic matter and activity of microbial population (Bastida et al. 2006).

In this study, protease, invertase, cellulase, urease, and acid phosphatase activities all significantly affected the rhizosphere in the six N. longibracteata forests. These rhizosphere effects may be due to the physiological activities of the roots under the influence of environmental stressors, which can cause the roots to produce large amounts of enzymes and release them into the rhizosphere soil. Meanwhile, microorganisms near the roots unceasingly secrete various enzymes to the surrounding soil root, resulting in very different enzyme activity profiles between the inner and outer rhizosphere (Yuan et al. 1997). These findings are consistent with these previous studies (Tscherko et al. 2004; Zhang et al. 2011), which suggests that the enzymes in the rhizosphere are more active than those in the bulk soil. However, Zhang et al. (2012) obtained that urease activity in the mineralization of soil N in rhizosphere is lower than that in the bulk soil, which may be due to the composition of the microbial communities. It was reported that soil enzymes are produced by specialized groups of microorganisms (Buée et al. 2009), which have competitive abilities for absorbable plant carbon.

On the other hand, rhizosphere effects may also be due to the physiological metabolic activities of diverse microbial populations that differ between rhizosphere zones (Garcia et al. 2005; Marschner et al. 2005; Wang et al. 2006; Phillips and Fahey 2008; Zhang et al. 2013). For example, higher acid phosphatase and protease activities in the rhizosphere versus bulk soil were observed by Marschner et al. (2005). Wang et al. (2006) found that rhizosphere soil had higher alkaline phosphatase activity than bulk soil. Phillips and Fahey (2008) reported that phosphatase enzyme activity was higher in rhizosphere than in the bulk soil of sugar maples (Acer saccharum Marshall) and northern red oak (Quercus rubra L.) at the Turkey Hill Plantation, New York (USA). Zhang et al. (2013) found that urease, catalase, invertase, acid phosphatase, and alkaline phosphatase activities in the rhizosphere soil were higher than bulk soil, while neutral phosphatase activity was much lower in the rhizosphere soils. Zhang et al. (2011) found a similar result in that urease and cellulase of rhizosphere soil was higher compared with that of the control soil.

The correlation analysis showed that the relationship among rhizosphere soil enzyme activities varied in the Tianbaoyan National Nature Reserve. There were significant linear correlations between phosphatase and protease activities and between acid phosphatase and cellulase activities in rhizosphere soil. In the study of Zhang et al. (2013), they found significant linear correlations between rhizosphere soil microbial abundances and enzyme activities. However, Ge et al. (2011) found the linear correlations between rhizosphere soil microbial abundances and enzyme activities could be due to the plant species, components of litter and the soil pathway of humus decomposition. Additional effort should be made to identify the microbial abundances to improve our understanding of rhizosphere soil microbial mechanisms.

Conclusion

Our study explored enzyme activities of rhizosphere and bulk soils of six N. longibracteata forests. We found that protease, invertase, cellulase, urease, and acid phosphatase activities in rhizosphere soil were significantly higher than those in bulk soils. Acid phosphatase activity showed significant rhizosphere effects in all six forests. Canonical correlation analysis of rhizosphere soil showed that acid phosphatase activity correlated significantly with protease activity (p < 0.01) and cellulase activity (p < 0.05). In bulk soil samples, there were significant linear correlations between protease and acid phosphatase activities (p < 0.05) and between invertase and cellulase activities (p < 0.01). Future studies would explore the relationship between microbial profiles and enzyme activities in rhizosphere soils. The first three PCs of the 13 chemical properties and enzyme activities accounted for 61.233, 15.422, and 9.434% of the total variances, respectively. Protease, invertase, acid phosphatase, total N, and cellulase had higher loadings on PC1, while total K and available K had higher loadings on PC2. PC3 was dominated by available P.