1. Understanding the effects of tree species and their mycorrhizal association on soil processes is critical for predicting the ecosystem consequences of species shifts owing to global change and forest management decisions. While it is well established that forests dominated by different mycorrhizal types can vary in how they cycle carbon (C), nitrogen (N) and phosphorus (P), the degree to which these patterns are driven by microbial-mediated enzyme activity (EA) and ecoenzymatic stoichiometry (ES) remain elusive.

2. Here, we synthesized the effects of mycorrhizal association on seven soil enzymes involved in microbial C, N and P acquisition and ES using data from 56 peer-reviewed papers.

3. We found that relative to soil in ectomycorrhizal (EcM) trees, soil in arbuscular mycorrhizal (AM) trees exhibited greater activity of some C acquisition enzymes (e.g., beta-glucosidase; BG) and higher ecoenzymatic ratios of BG/NAG (N-acetyl-glucosaminidase) and BG/AP (acid phosphatase). These results supported that AM trees had rapid C and nutrient turnover rates, inorganic nutrient economics and high soil microbial C limitation. We also found evidence for an organic nutrient economy and greater soil microbial demand for nutrients in EcM trees compared to AM trees. In addition, the effect of mycorrhizal association on the activity of certain soil enzymes and enzymatic stoichiometry (i.e., BG and BG/NAG ratio) appeared to be associated with the differences in soil pH, phylogenetic group (i.e., conifers and broadleaves) and leaf habit (i.e., evergreen and deciduous) between AM and EcM trees.

4. The results from the global meta-analysis suggested that soil EA and ES appear to play critical roles in shaping the differences in the nutrient economy between AM and EcM tree species, but leaf morphology and soil conditions should be considered in evaluations of soil processes in forests of different mycorrhizal associations. Given that most of the studies in the database were from the temperate and subtropical regions, further research in other biomes is needed to elucidate the underlying mechanisms driving the mycorrhizal effect at the global scale.

We searched for peer-reviewed journal articles on the topic of tree species and soil enzyme activity spanning 1970–2021 using the Web of Science (http://apps.webofknowledge.com/) and the China National Knowledge Infrastructure Database (CNKI; http://www.cnki.net/). The following terms were used in English and Chinese (for CNKI) for the search: (Mycorrhizal* OR "tree species" OR plantation* OR afforestation) AND (enzyme*). The gathered publications were further screened based on four criteria: (1) co-occurring AM and ECM forests must grow on the same site with similar edaphic and climatic conditions; (2) AM and ECM forests must be of the same age, and at least 10 years old; (3) field studies must have been conducted in monoculture stands (i.e. composed by a single tree species) or natural forests (coverage of dominant species > 75% or the author clarified the dominant mycorrhizal types); and (4) at least one of the enzymes listed in Table 1 must be reported.

A total of 56 studies containing 286 paired observations were identified (Fig. 1, Appendix 1, and Appendix 2). We first extracted the data of various soil enzymes, including BG, invertase (INV), NAG, urease (URE), AP, PPO and POD (Table 1) as they release C or nutrients from organic compounds. Data were directly obtained from tables or extracted from figures using GetData (version 2.25). We also contacted the corresponding authors, if the original enzyme data was not reported. The classification of mycorrhizal fungal types was based on the newly published database (Soudzilovskaia et al., 2020) and was confirmed by using the diagnosed database from previous studies (Brundrett & Tedersoo, 2020; M. Brundrett & Tedersoo, 2019; Tedersoo & Brundrett, 2017). To determine deciduousness (i.e., evergreen or deciduous) and phylogeny (broadleaf or conifer) of each tree species, we extracted directly from the literature if available or searched the relevant database (i.e., Flora Reipublicae Popularis Sinicae and The World Flora) that reported such information. The enzymatic ratios BG/NAG, BG/AP, and NAG/AP were calculated to indicate C/N, C/P and N/P demand of the microbes, respectively. The enzymatic ratios were only calculated if the enzyme measurements were based on the same colorimetric substrates (e.g., p-nitrophenol (pNP)) or fluorescent substrates (i.e., 4-methylumbelliferone (MUB)). Apart from soil enzyme activity, soil physicochemical properties (e.g., soil pH, soil organic C and C/N) in the same layer were also collected. Furthermore, site-specific information has also been collected, including geophysical variables (latitude, longitude, and altitude), climatic variables (mean annual temperature (MAT), mean annual precipitation (MAP)), Latin names of tree species, forest stand age and soil depth sampled for EA analysis. If the MAT and MAP were not directly provided from the papers, the site information (latitude and longitude) was employed to extract them from a global database (http://www.worldclim.org/).