Soil bacterial community responses to altered precipitation and temperature regimes in an old field grassland are mediated by plants

Soil bacterial community responses to altered precipitation and temperature regimes in an old... Abstract The structure and function of soil microbiomes often change in response to experimental climate manipulations, suggesting an important role in ecosystem feedbacks. However, it is difficult to know if microbes are responding directly to environmental changes or are more strongly impacted by plant responses. We investigated soil microbial responses to precipitation and temperature manipulations at the Boston-Area Climate Experiment in Massachusetts, USA, in both vegetated and bare plots to parse direct vs. plant-mediated responses to multi-factor climate change. We assessed the bacterial community in vegetated soils in 2009, two years after the experiment was initiated, and bacterial and fungal community in vegetated and bare soils in 2011. The bacterial community structure was significantly changed by the treatments in vegetated soils. However, such changes in the bacterial community across the treatments were absent in the 2011 bare soils. These results suggest that the bacterial communities in vegetated soils were structured via plant community shifts in response to the abiotic manipulations. Co-variation between bacterial community structure and temperature sensitivities and stoichiometry of potential enzyme activities in the 2011 vegetated soils suggested a link between bacterial community structure and ecosystem function. This study emphasizes the importance of plant-soil-microbial interactions in mediating responses to future climate change. Boston-Area Climate Experiment, precipitation, temperature, bacterial and fungal community structure, potential enzyme activity INTRODUCTION Soil microbial communities appear to be sensitive to climatic changes that are characterized by increasing temperature and intensified precipitation regimes in many parts of the world (Min et al.2011; Cai et al.2014). However, it is unclear whether responses of microbiome structure and function in experimental climate change are direct or are mediated by simultaneous plant community shifts in response to the abiotic changes (Legay et al.2014). It is well established that abiotic factors such as pH (Fierer and Jackson 2006; Lauber et al.2009), temperature (Feng and Simpson 2009) and moisture availability (Brockett, Prescott and Grayston 2012) are major drivers of microbial community structure. Plant community shifts in response to climate change factors have been observed in natural systems (Kelly and Goulden 2008; Chen et al.2011) and in experimental manipulations of precipitation (Kardol et al.2010; Yang et al.2011) and temperature (Cowles et al.2016; Mueller et al.2016). Such shifts in vascular plant communities can shape the microbial community structure in the rhizosphere (Philippot et al.2013). Therefore, both abiotic factors and plant community shifts could structure soil microbial communities under changing climates. Changes in microbial community structure in response to the environmental changes could also alter biogeochemical functioning. Soil microbes, including bacteria and fungi, play vital roles in biogeochemical cycling of carbon (C) and nutrients such as nitrogen (N) and phosphorus (P) (Falkowski, Fenchel and Delong 2008). These roles include serving as the base of the soil food web (Moore et al.2004), helping form and stabilize soil organic matter (Averill, Turner and Finzi 2014; Cotrufo et al.2015), and mineralizing of soil organic matter, which makes available nutrients for plants and microbes themselves (Romaní et al.2006). As providers of the initial, rate-limiting step of organic matter mineralization, soil microbes produce extracellular enzymes to depolymerize and solubilize large organic compounds (Wallenstein and Weintraub 2008). Characteristics of extracellular enzymes, such as their temperature sensitivity (Alster et al.2016a), and stoichiometry among C and nutrients (Makino et al.2003), vary among different soil microbes. Therefore, shifts in soil microbial communities potentially affect soil enzymatic properties (Blagodatskaya et al.2016), even though it is possible that significantly different microbial communities can result in similar functions (Gibbons et al. 2017; Louca et al.2017). Although co-variation between microbial community and biogeochemical processes, such as nitrifiers and nitrification (e.g. Rudisill, Turco and Hoagland 2016) and methanotrophs and methane oxidation (e.g. Levine et al.2011, Judd et al.2016), has been reported, such relationships between microbial community structure and soil enzyme properties have not been well explored (Bier et al.2015), In this study, we investigated how manipulations of precipitation and temperature influenced microbial community structure and biogeochemical properties in soils at the Boston-Area Climate Experiment (BACE) in a mesic old-field community in Massachusetts, USA. We assessed the community structure of bacteria in vegetated soils in June 2009, two years after the experiment was initiated, and that of bacteria and fungi in both vegetated and bare soils in July 2011 during the growing season, using Roche 454 sequencing. We hypothesized that (i) microbial populations in the vegetated soils were more affected by the precipitation and temperature treatments than those in bare soils, and (ii) bacterial and fungal communities responded to the treatments differently as the two taxa possess distinctive physiological differences such as biomass turnover rates (Rousk and Bååth 2011) and carbon use efficiency (Six et al.2006). We measured soil properties including water content, organic C and total N, and substrate-induced respiration as an index for microbial biomass. To investigate potential effects of microbial community structure on biogeochemical processes, we quantified activities of seven extracellular enzymes involved in hydrolysis of C and N compounds and phosphate, in vegetated soils collected in 2011. The enzyme assay was conducted at four different temperatures in the lab to assess temperature sensitivities of each of the seven enzymes. We explored the relationship between microbial community structure and enzyme stoichiometry and temperature sensitivity via a multivariate analysis. MATERIALS AND METHODS Study site The study was conducted at the BACE site, located in an old-field community in Waltham, Massachusetts, USA (42°23΄ 3″N, 71° 12΄ 52″W). Mean annual temperature and precipitation are 9.5°C and 1194 mm, respectively (Hoeppner and Dukes 2012). Three levels of precipitation and four levels of temperature treatments had been applied to 2 m by 2 m plots in a randomized, complete block, split-plot design (N = 3) since 2007 and 2008, respectively. This experimental design resulted in a total of 36 plots, which were the study and statistical units. The three precipitation treatments were drought (50% of ambient precipitation, year-round), ambient, and wet (150% of ambient during the growing season). The four temperature treatments were control, low (+∼1°C), medium (+∼2.7°C) and high (+∼4°C) warming using ceramic infrared heaters (Hoeppner and Dukes 2012). The study site had a loamy top soil (Suseela et al.2013). At the onset of the experiment in 2007, the vegetation was a mixture of native and introduced forbs and grasses, where C3 species, perennials and grasses were dominant over C4 species, annuals, and forbs, respectively (Hoeppner and Dukes 2012). By the growing season of 2010, the precipitation and temperature treatments had affected the plant community, where C4 grasses and other annual and biennial species entered the C3 perennial-dominated community in the ambient and wet treatments and the warmed treatments (Hoeppner and Dukes 2012). However, the precipitation and warming treatments did not significantly alter total plant production (Hoeppner and Dukes 2012). Soil sampling, processing and analyses Surface soils to 5 cm in depth were collected from vegetated areas in June 2009, and, in July 2011, from vegetated areas as well as bare surfaces within PVC collars (25 and 30 cm in diameter and depth, respectively) that excluded growth of vascular plants (Suseela et al.2012), in all 36 plots. The soil samples were transported on ice overnight to Natural Resource Ecology Laboratory, Colorado State University, Colorado, USA. The soils were sieved with a 2 mm screen and kept at 4°C for subsequent measurements. Each soil sample was quantified for soil water content (SWC), soil organic carbon (SOC), total nitrogen (N) and substrate-induced respiration rate. SWCs were measured by drying the soils at 105°C for 48 h. For SOC and total N analyses, soils were first dried out at 60°C, and ground using a Brinkmann Retsch mill (Haan, Germany), followed by quantification with a LECO TruSpec® (Leco Corporation, St. Joseph, Michigan, USA). Substrate-induced respiration rates were measured using MicroRespTM (Aberdeen, UK, Campbell et al.2003) in the same manner described in detail by Steinweg et al. (2013). Briefly, each sieved soil was amended with deionized water to 55% of the water holding capacity, and added to a deep 96-well plate with three technical replicates. Following incubation at 4°C for 18 h, 25 μL of 1 M glucose solution was added to each well and incubated at 15°C for 6 h. The same volume of deionized water was added to each soil in another plate. Differences in quantities of CO2 produced between the glucose- and deionized water-amended sub-samples during the 6-h incubation were calculated as substrate induced respiration. Enzyme assays Potential enzyme activities were assessed for the vegetated soils collected in 2011 using a fluorometric method (Saiya-Cork, Sinsabaugh and Zak 2002) modified by Steinweg et al. (2012) in the same manner described in details by Koyama et al. (2013). In total, seven enzymes were assessed: four enzymes to hydrolyze C-rich substrates, including β-glucosidase (BG), cellobiohydrolase (CB), xylosidase (XYL) and α-glucosidase (AG), two enzymes to hydrolyze N-rich substrates including n-acetyl-glucosaminidase (NAG) and leucine-amino-peptidase (LAP), and phosphatase (PHOS). Briefly, each soil sub-sample of 2.75 g was homogenized in 91 mL of 50 mM sodium acetate (pH 5.5) in a blender. An 800 μL aliquot of the slurry was pipetted into each of seven wells in a column of a 96 deep-well plate, and mixed with 200 μL solution of each of the seven substrates. Substrate concentrations had been pre-determined so that they were not completely consumed during incubation. Reference standards were prepared in a similar manner; 200 μL of fluorescent standard, ranging from zero to 200 μM, were mixed with an aliquot of 800 μL slurry for each soil. Two types of standards were used; 7-amino-4-methylcoumarin (MUC) for LAP and 4-methylumbelliferone (MUB) for the other enzymes. Four plates were prepared for each soil sample to assess potential enzyme activities at four different temperatures (5°C, 15°C, 25°C and 35°C) that were used to assess temperature sensitivity of the activities. The incubation duration at 5°C, 15°C, 25°C and 35°C were 23, 6, 3 and 1.5 h, respectively. After incubation, the plates were centrifuged at 350 g for 3 min, and 250 μL of supernatant from each well was placed into the corresponding well of a black 96-well plate. Fluorescence was measured with 365 and 450 nm in wavelengths for excitation and emission, respectively, using a Tecan Infinite M500 (Tecan Group, Ltd., Männedorf, Switzerland). Linear standard curves, obtained using the MUC or MUB standards, were used to calculate potential enzyme activity for each enzyme and sample. To assess temperature sensitivity of the enzymes, activation energy was calculated using the potential enzyme activities assayed at the four temperatures using the Arrhenius equation (Arrhenius 1889) described in details in Koyama et al. (2013). DNA extraction, PCR and 454 sequencing Subsets of the soil samples were analyzed for the microbial community structure using Roche 454 sequencing. In 2009, soils from nine plots with the control temperature treatment across the three precipitation treatments and three plots with the ambient precipitation and high warming were selected. In 2011, soils from all the plots with the control and high warming treatments across the three precipitation treatments were selected. The selected samples of 2009 and 2011 were processed in the same manner described in details by Evans and Wallenstein (2012) and Koyama et al. (2014), respectively. Briefly, genomic DNA was extracted from each of 0.25 g sub-sample collected from the control and high-temperature treatments using MoBio PowerSoil DNA extraction kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA) and stored at −80°C until further processing. The small subunit of 16S and 18S rRNA genes was amplified for each sample using paired primers of F515/R806 (Bates et al.2010) and SSU817R/SSU1196 (Borneman and Hartin, 2000), respectively. The primers were modified for the 454 sequencing platform (Rousk et al.2010). For the samples collected in 2009, PCR was performed using 50 μL assays; 0.5 μL (10 μM) of each primer, 3 μL (5 ng μL−1) of template DNA, 5 μL of BSA (5 ng μL−1), 18.5 μL of PCR-grade water and 22.5 μL of Platinum PCR SuperMix (Invitrogen, Carlsbad, CA, USA). The PCR thermal profile consisted of an initial denaturation at 94°C for 3 min, followed by 35 cycles of 95°C for 45 s, 50°C for 30 s and 72°C for 90 s with a final extension of 10 min at 72°C. For the samples collected in 2011, PCR was performed using 25 μL assays; 1.25 μL (10 μM) of each primer, 1 μL (1 ng μL−1) of template DNA, 1.25 μL of BSA (10 ng μL−1), 8.5 μL of PCR-grade water, and 12.5 μL of KAPA2G Fast Multiplex Mix (Kapa Biosystems, Woburn, MA, USA). The thermal profile consisted of an initial denaturation at 95°C for 3 min, followed by 30 cycles of 95°C for 10 s, 50°C for 10 s and 72°C for 1 s with a final extension of 10 min at 72°C. Amplicons were evaluated for amplification and lengths by agarose gel electrophoresis, and purified using the UltraClean® PCR Clean-UP Kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA). Appropriate quantities of purified amplicons were pooled and sequenced on a Roche 454 FLX sequencer at Selah Genomics (Greenville, SC, USA). Sequence data processing Sequences were processed via the QIIME 1.9 toolkit (Caporaso et al.2010a). Operational taxonomic units (OTUs) were determined at the ≥97% in similarity of the sequences, and assigned taxonomy via the Ribosomal Database Project (Cole et al.2009) and NCBI BLAST (Johnson et al.2008) for bacteria and fungi, respectively. After chimeric sequences and singletons were removed, total numbers of sequences of the 2009 bacteria, 2011 bacteria and 2011 fungi were 67 706 (ranging from 2513 to 7620 sequences per sample), 63 025 (660 to 8337), and 102 588 (1020 to 4116), respectively. The bacterial sequences collected in 2009 and 2011 were rarefied at 2513 and 660 per sample, respectively, and the fungal sequence at 1020 per sample for downstream analyses. Sequences in each data set were aligned via PyNAST (Caporaso et al.2010b) to build a phylogenetic tree using FastTree (Price, Dehal and Arkin 2009). The downstream analyses included UniFrac (Hamady, Lozupone and Knight 2010) in QIIME, and distant-based redundancy analysis (dbRDA, Legendre and Anderson 1999) using the vegan package in R (R Development Core Team 2015). The sequences were deposited to the MG-RAST server (http://metagenomics.anl.gov/) and are available to the public (accession numbers from 4735126.3 to 4758399.3 under ‘BACE_soil_microbes’). Statistical analyses All the computations were carried out using R (R Development Core Team 2015). Mixed-effect ANOVAs in the lme4 package were employed, with precipitation (i.e. drought, ambient and wet) and temperature (i.e. control, low, medium and high) as fixed effects, and blocks as a random effect. A significance level of P ≤ 0.05 was considered significant. RESULTS Soil properties Both precipitation and temperature treatments significantly altered some soil properties (Fig. 1). Overall, SWC was lowest in the drought and high-temperature treatments (Fig. 1). In 2009, the differences in SWC across the precipitation treatments were widened as the higher heat was applied, supported by a precipitation × temperature interaction effect (P ≤ 0.001, Fig. 1). In the 2011 soils, the drought (P ≤ 0.001, Table 1) and higher temperature treatments (P ≤ 0.001, Table 1) reduced SWC (Fig. 1). The soils under vegetation had lower SWC than bare soils (P ≤ 0.001, Table 1, Fig. 1). Figure 1. View largeDownload slide Properties of soils collected in 2009 and 2011. The soils collected in 2009 were from vegetated areas only. Soils collected in 2011 were from two cover types (vegetated and bare soils). Results of mixed-effect ANOVAs for the 2011 data with the two cover types combined are shown in Table 1. Results of mixed-effect ANOVAs for each cover type are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Figure 1. View largeDownload slide Properties of soils collected in 2009 and 2011. The soils collected in 2009 were from vegetated areas only. Soils collected in 2011 were from two cover types (vegetated and bare soils). Results of mixed-effect ANOVAs for the 2011 data with the two cover types combined are shown in Table 1. Results of mixed-effect ANOVAs for each cover type are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Table 1. Results of mixed-effect ANOVAs for properties of soils collected in June 2011. Soil samples from the two cover types (vegetated and bare soils) were pooled for the analyses. P-values ≤ 0.05 are in bold. Response variables  Independent variables  F values  P values  SWC  Temp  F3,22 = 20.898  P < 0.001    Precip  F2,22 = 24.89  P < 0.001    Cover  F1,24 = 122.294  P < 0.001    Temp × Precip  F6,22 = 1.056  P = 0.418    Temp × Cover  F3,24 = 0.647  P = 0.593    Precip × Cover  F2,24 = 0.993  P = 0.385    Temp × Precip × Cover  F6,24 = 1.188  P = 0.346  SOC  Temp  F3,22 = 0.596  P = 0.625    Precip  F2,22 = 5.825  P = 0.009    Cover  F1,24 = 1.676  P = 0.208    Temp × Precip  F6,22 = 0.572  P = 0.748    Temp × Cover  F3,24 = 0.379  P = 0.769    Precip × Cover  F2,24 = 1.144  P = 0.335    Temp × Precip × Cover  F6,24 = 1.193  P = 0.344  Total N  Temp  F3,22 = 0.321  P = 0.810    Precip  F2,22 = 4.065  P = 0.031    Cover  F1,24 < 0.001  P = 0.987    Temp × Precip  F6,22 = 0.551  P = 0.764    Temp × Cover  F3,24 = 0.384  P = 0.765    Precip × Cover  F2,24 = 0.922  P = 0.412    Temp × Precip × Cover  F6,24 = 1.224  P = 0.329  C:N ratio  Temp  F3,22 = 0.86  P = 0.478    Precip  F2,22 = 7.41  P = 0.004    Cover  F1,24 = 35.35  P < 0.001    Temp × Precip  F6,22 = 0.81  P = 0.570    Temp × Cover  F3,24 = 1.19  P = 0.333    Precip × Cover  F2,24 = 8.13  P = 0.002    Temp × Precip × Cover  F6,24 = 1.69  P = 0.167  SIR  Temp  F3,22 = 4.916  P = 0.009    Precip  F2,22 = 18.952  P < 0.001    Cover  F1,24 = 716.886  P < 0.001    Temp × Precip  F6,22 = 1.109  P = 0.389    Temp × Cover  F3,24 = 1.78  P = 0.178    Precip × Cover  F2,24 = 8.087  P = 0.002    Temp × Precip × Cover  F6,24 = 0.419  P = 0.859  Response variables  Independent variables  F values  P values  SWC  Temp  F3,22 = 20.898  P < 0.001    Precip  F2,22 = 24.89  P < 0.001    Cover  F1,24 = 122.294  P < 0.001    Temp × Precip  F6,22 = 1.056  P = 0.418    Temp × Cover  F3,24 = 0.647  P = 0.593    Precip × Cover  F2,24 = 0.993  P = 0.385    Temp × Precip × Cover  F6,24 = 1.188  P = 0.346  SOC  Temp  F3,22 = 0.596  P = 0.625    Precip  F2,22 = 5.825  P = 0.009    Cover  F1,24 = 1.676  P = 0.208    Temp × Precip  F6,22 = 0.572  P = 0.748    Temp × Cover  F3,24 = 0.379  P = 0.769    Precip × Cover  F2,24 = 1.144  P = 0.335    Temp × Precip × Cover  F6,24 = 1.193  P = 0.344  Total N  Temp  F3,22 = 0.321  P = 0.810    Precip  F2,22 = 4.065  P = 0.031    Cover  F1,24 < 0.001  P = 0.987    Temp × Precip  F6,22 = 0.551  P = 0.764    Temp × Cover  F3,24 = 0.384  P = 0.765    Precip × Cover  F2,24 = 0.922  P = 0.412    Temp × Precip × Cover  F6,24 = 1.224  P = 0.329  C:N ratio  Temp  F3,22 = 0.86  P = 0.478    Precip  F2,22 = 7.41  P = 0.004    Cover  F1,24 = 35.35  P < 0.001    Temp × Precip  F6,22 = 0.81  P = 0.570    Temp × Cover  F3,24 = 1.19  P = 0.333    Precip × Cover  F2,24 = 8.13  P = 0.002    Temp × Precip × Cover  F6,24 = 1.69  P = 0.167  SIR  Temp  F3,22 = 4.916  P = 0.009    Precip  F2,22 = 18.952  P < 0.001    Cover  F1,24 = 716.886  P < 0.001    Temp × Precip  F6,22 = 1.109  P = 0.389    Temp × Cover  F3,24 = 1.78  P = 0.178    Precip × Cover  F2,24 = 8.087  P = 0.002    Temp × Precip × Cover  F6,24 = 0.419  P = 0.859  View Large Contents of SOC were affected by the precipitation treatments. The 2009 vegetated soils had higher SOC content in the drought than the control or wet treatment (P = 0.038, Fig. 1). Overall, a similar trend was found in the soils collected in July 2011 (P = 0.009, Table1, Fig. 1). However, this trend was not significant when only the vegetated soils were examined (P = 0.186). On the other hand, the bare soils in the drought and wet treatments tended to have higher SOC contents than controls (P = 0.011). No treatment significantly altered total N contents in the 2009 vegetated soils (Fig. 1). Overall, total N contents in soils collected in 2011 showed a trend similar to SOC in the same soil samples (Fig. 1) where the precipitation altered total N content (P = 0.009, Table 1). A significant precipitation effect was not detected in the vegetated soils (P = 0.181, Fig. 1), but was found in the bare soils (P = 0.040, Fig. 1). No treatment significantly altered C:N ratios of the 2009 vegetated soils (Fig. 1). The precipitation treatment significantly altered C:N ratio of the 2011 bare soils, but this was not the case for the 2011 vegetated soils (Fig. 1). This was supported by an interaction effect between the precipitation and cover type for the 2011 soils (P = 0.002, Table 1). When the two cover types were compared, C:N ratios were higher in the bare than vegetated soils (P ≤ 0.001, Table 1). In the bare soils, C:N ratios of soils under the drought treatment were highest, followed by those of control and wet precipitation treatments (Fig. 1). No treatment significantly altered SIR in the 2009 vegetated soils (Fig. 1). On the other hand, both temperature and precipitation treatments affected SIR in the 2011 soils (Fig. 1). Soils in the drought treatment showed lowest SIR and the wet treatment showed the highest (P ≤ 0.001, Table 1), and this trend was more pronounced in vegetated than bare soils (Fig. 1). This is supported by a precipitation × cover type effect (P = 0.002, Table 1). The vegetated soils had higher SIR than bare soils (P ≤ 0.001, Table 1). In both types of soils, higher temperature treatments decreased SIR (P = 0.009, Table 1). When SIR data were analyzed for each cover type, both precipitation (P = 0.005 and < 0.001 in vegetated and bare soils, respectively) and temperature treatments (P = 0.023 and 0.021 in vegetated and bare soils, respectively) were significant (Fig. 1). Microbial community composition Bacterial community composition differed among the precipitation treatments in the 2009 vegetated soils, as demonstrated by a UniFrac analysis (P ≤ 0.001 in PC1, Fig. 2). The PC1 values, which explained 49.1% of the variation, co-varied with precipitation quantities, where the drought treatment had the lowest, and the wet treatment the highest values (Fig. 1). When the control and high-temperature treatments were compared for the soils in the ambient precipitation via paired Student t-tests, there were significant differences in both PC1 and PC2 (P = 0.020 and 0.011, respectively). Figure 2. View largeDownload slide Results of UniFrac analyses for bacterial and fungal community composition. Symbols and error bars represent means and standard errors, respectively. Soils collected in 2009 were from vegetated areas subjected to four treatments, and processed for bacterial community composition only. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat) and processed for bacterial and fungal community composition. The UniFrac analyses for the 2011 data were conducted with the two cover types together, but the results of the two cover types were plotted separately to help distinguish the differences. ANOVAs were conducted for PC values of the 2009 soils with the control temperature treatment. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data are shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Figure 2. View largeDownload slide Results of UniFrac analyses for bacterial and fungal community composition. Symbols and error bars represent means and standard errors, respectively. Soils collected in 2009 were from vegetated areas subjected to four treatments, and processed for bacterial community composition only. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat) and processed for bacterial and fungal community composition. The UniFrac analyses for the 2011 data were conducted with the two cover types together, but the results of the two cover types were plotted separately to help distinguish the differences. ANOVAs were conducted for PC values of the 2009 soils with the control temperature treatment. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data are shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. A similar trend was observed for bacterial community composition in the 2011 vegetated soils (Fig. 2). Values of the PC1 co-varied with the precipitation as well as temperature gradients (P < 0.001, and = 0.002, respectively, Fig. 2) when only the PC1 values of the vegetated soils were processed via an ANOVA. The PC2 values were different among the precipitation treatments, where the soils under the ambient treatment had higher scores than the manipulated precipitation treatments (P = 0.026). In contrast to the vegetated soils, there were no precipitation or temperature treatment effects in the bare soils in the PC1 or PC2 values (P ≥ 0.097, Fig. 2). Such contrasting responses of bacterial community composition between the two cover types were supported by a precipitation × cover types effect for the PC1 values (P = 0.002, Table 2). In PC2, most of the values of vegetated and bare soils were positive and negative, respectively (Fig. 2) supported by a significant main cover type effect for the PC2 values (P < 0.001, Table 1). The PC1 and PC2 explained 18.7% and 12.0% of variability in bacterial community structure in the UniFrac analysis (Fig. 2). Table 2. Results of mixed-effect ANOVAs for sample scores of two primary axes resuling from multivariate analyses of bacterial and fungal community composition. Microbial data were obtained from soil samples from the two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat). P-values ≤ 0.05 are shown bold. Response variables  Independent variables  F values  P values  Bacterial weighted UniFrac PC 1  Temp  F1,19 = 5.535  P = 0.030    Precip  F2,19 = 5.555  P = 0.013    Cover  F1,19 = 0.471  P = 0.501    Temp × Precip  F2,19 = 0.146  P = 0.865    Temp × Cover  F1,19 = 1.136  P = 0.300    Precip × Cover  F2,19 = 9.165  P = 0.002    Temp × Precip × Cover  F2,19 = 0.457  P = 0.640  Bacterial weighted UniFrac PC 2  Temp  F1,19 = 6.494  P = 0.020    Precip  F2,19 = 7.693  P = 0.004    Cover  F1,19 = 81.616  P < 0.001    Temp × Precip  F2,19 = 2.54  P = 0.105    Temp × Cover  F1,19 = 0.005  P = 0.944    Precip × Cover  F2,19 = 2.128  P = 0.147    Temp × Precip × Cover  F2,19 = 1.062  P = 0.365  Fungal weighted UniFrac PC 1  Temp  F1,22 = 0.001  P = 0.979    Precip  F2,22 = 0.101  P = 0.905    Cover  F1,22 = 0.261  P = 0.614    Temp × Precip  F2,22 = 0.787  P = 0.468    Temp × Cover  F1,22 = 0.784  P = 0.385    Precip × Cover  F2,22 = 0.485  P = 0.622    Temp × Precip × Cover  F2,22 = 1.143  P = 0.337  Fungal weighted UniFrac PC 2  Temp  F1,22 = 12.535  P = 0.002    Precip  F2,22 = 0.78  P = 0.471    Cover  F1,22 = 0.059  P = 0.811    Temp × Precip  F2,22 = 2.042  P = 0.154    Temp × Cover  F1,22 = 0.141  P = 0.711    Precip × Cover  F2,22 = 5.051  P = 0.016    Temp × Precip × Cover  F2,22 = 0.749  P = 0.485  Bacterial dbRDA Axis 1  Temp  F1,19 = 10.074  P = 0.005    Precip  F2,19 = 7.016  P = 0.005    Cover  F1,19 = 6.128  P = 0.023    Temp × Precip  F2,19 = 0.069  P = 0.933    Temp × Cover  F1,19 = 0.547  P = 0.468    Precip × Cover  F2,19 = 10.373  P = 0.001    Temp × Precip × Cover  F2,19 = 0.28  P = 0.759  Bacterial dbRDA Axis 2  Temp  F1,19 = 0.032  P = 0.860    Precip  F2,19 = 1.377  P = 0.277    Cover  F1,19 = 19.613  P < 0.001    Temp × Precip  F2,19 = 3.217  P = 0.063    Temp × Cover  F1,19 = 1.135  P = 0.300    Precip × Cover  F2,19 = 9.765  P = 0.001    Temp × Precip × Cover  F2,91 = 3.932  P = 0.037  Response variables  Independent variables  F values  P values  Bacterial weighted UniFrac PC 1  Temp  F1,19 = 5.535  P = 0.030    Precip  F2,19 = 5.555  P = 0.013    Cover  F1,19 = 0.471  P = 0.501    Temp × Precip  F2,19 = 0.146  P = 0.865    Temp × Cover  F1,19 = 1.136  P = 0.300    Precip × Cover  F2,19 = 9.165  P = 0.002    Temp × Precip × Cover  F2,19 = 0.457  P = 0.640  Bacterial weighted UniFrac PC 2  Temp  F1,19 = 6.494  P = 0.020    Precip  F2,19 = 7.693  P = 0.004    Cover  F1,19 = 81.616  P < 0.001    Temp × Precip  F2,19 = 2.54  P = 0.105    Temp × Cover  F1,19 = 0.005  P = 0.944    Precip × Cover  F2,19 = 2.128  P = 0.147    Temp × Precip × Cover  F2,19 = 1.062  P = 0.365  Fungal weighted UniFrac PC 1  Temp  F1,22 = 0.001  P = 0.979    Precip  F2,22 = 0.101  P = 0.905    Cover  F1,22 = 0.261  P = 0.614    Temp × Precip  F2,22 = 0.787  P = 0.468    Temp × Cover  F1,22 = 0.784  P = 0.385    Precip × Cover  F2,22 = 0.485  P = 0.622    Temp × Precip × Cover  F2,22 = 1.143  P = 0.337  Fungal weighted UniFrac PC 2  Temp  F1,22 = 12.535  P = 0.002    Precip  F2,22 = 0.78  P = 0.471    Cover  F1,22 = 0.059  P = 0.811    Temp × Precip  F2,22 = 2.042  P = 0.154    Temp × Cover  F1,22 = 0.141  P = 0.711    Precip × Cover  F2,22 = 5.051  P = 0.016    Temp × Precip × Cover  F2,22 = 0.749  P = 0.485  Bacterial dbRDA Axis 1  Temp  F1,19 = 10.074  P = 0.005    Precip  F2,19 = 7.016  P = 0.005    Cover  F1,19 = 6.128  P = 0.023    Temp × Precip  F2,19 = 0.069  P = 0.933    Temp × Cover  F1,19 = 0.547  P = 0.468    Precip × Cover  F2,19 = 10.373  P = 0.001    Temp × Precip × Cover  F2,19 = 0.28  P = 0.759  Bacterial dbRDA Axis 2  Temp  F1,19 = 0.032  P = 0.860    Precip  F2,19 = 1.377  P = 0.277    Cover  F1,19 = 19.613  P < 0.001    Temp × Precip  F2,19 = 3.217  P = 0.063    Temp × Cover  F1,19 = 1.135  P = 0.300    Precip × Cover  F2,19 = 9.765  P = 0.001    Temp × Precip × Cover  F2,91 = 3.932  P = 0.037  View Large When the bacterial community composition was further examined at a higher taxonomic level using dbRDA, similar trends were found (Fig. 3). The axis 1 values of the 2009 soils collected under vegetation, which explained 53.9% of the variability, co-varied with water availability (P < 0.001, Fig. 3). The primary driver of the co-variation was due to the dominance of Actinobacteria in the drought treatment (Fig. 3; Fig. S1, Supporting Information). Along the axis 2, the drought and wet treatments had lower values than the ambient treatment (P = 0.003, Fig. 3). Figure 3. View largeDownload slide Results of distance-based redundancy analyses (dbRDA) for the composition of bacterial taxa in soils collected in 2009 and 2011. Soils collected in 2009 were from vegetated areas subjected to four treatments. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat). The dbRDA for the 2011 data were conducted with the two cover types combined, but the results of the two cover types were plotted separately to help distinguish the differences. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data were shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Proteo; Proteobacteria, Acido: Acidobacteria. Figure 3. View largeDownload slide Results of distance-based redundancy analyses (dbRDA) for the composition of bacterial taxa in soils collected in 2009 and 2011. Soils collected in 2009 were from vegetated areas subjected to four treatments. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat). The dbRDA for the 2011 data were conducted with the two cover types combined, but the results of the two cover types were plotted separately to help distinguish the differences. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data were shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Proteo; Proteobacteria, Acido: Acidobacteria. Bacterial community composition in the 2011 vegetated soils via dbRDA showed a similar trend to the UniFrac analysis (Fig. 3). The axis 1 values co-varied with the precipitation as well as temperature gradients supported by significant precipitation and temperature effects (P < 0.001, Fig. 3) when only the axis 1 values of the vegetated soils were analyzed via an ANOVA. This co-variation was primarily driven by the dominance of Actinobacteria in the drought and high-temperature treatments (Figs 3 and 4). In contrast to the vegetated soils, there were no significant differences among the precipitation or temperature treatments in the bare soils in the axis 1 values (P ≥ 0.189, Fig. 3). Such contrasting responses of bacterial community composition between the two cover types were supported by a precipitation × cover type effect as well as a main cover type effect for the axis 1 values (P = 0.001 and 0.023, respectively, Table 2). The axis 2 values of the vegetated soils were different among the precipitation treatments, where the soils under the ambient treatment had higher scores than the manipulated precipitation treatments (P = 0.026, Fig. 3). The axis 2 values of the bare soils indicated the two treatments significantly altered relative abundances of bacterial taxa, supported by temperature × precipitation and main precipitation effects (P = 0.05 and 0.036, respectively, Fig. 3). Overall, the vegetated soils had higher values than the bare soils along axis 2 (Fig. 3). This difference was supported by precipitation × cover type and main cover type effects (P = 0.001 and < 0.001, respectively, Table 1). The axes 1 and 2, respectively, explained 32.9% and 8.3% of variability in the bacterial community structure in the dbRDA analysis for the 2011 soils (Fig. 3). Figure 4. View largeDownload slide Mean relative abundances of major bacterial taxa in the vegetated and bare soils collected in 2011. Drought, ambient and wet represent three levels of the precipitation treatments. C and H represent the control and high-temperature treatments, respectively. Figure 4. View largeDownload slide Mean relative abundances of major bacterial taxa in the vegetated and bare soils collected in 2011. Drought, ambient and wet represent three levels of the precipitation treatments. C and H represent the control and high-temperature treatments, respectively. The fungal community structure was less affected by the treatments relative to that of bacteria (Fig. 2). When both cover types were analyzed together, there was no significant treatment effect on PC1, which explained 48.9% of variability (Table 2, Fig. 2). In PC2, which explained 11.9% of variability, the temperature treatment was significant (P = 0.002, Table 2) where the values of control tended to be higher than those of the high-temperature treatment when paired by the precipitation treatments for each cover type (Fig. 2). There was a precipitation × cover type effect for PC2 (P = 0.016, Table 2). This was likely caused by the significant precipitation effect in the vegetated soils but a lack of this trend in the bare soils when the cover types were analyzed separately (Fig. 2). Due to the lack of clear responses to the treatments and low taxonomic resolution (Figs S2 and S3, Supporting Information), fungal relative abundances at higher taxonomic ranks were not analyzed further. Enzyme assay Overall, four enzymes, BG, CB, LAP and PHOS, had higher potential activities when more precipitation was applied and had lower activities when increasing heat was applied in the 2011 vegetated soils (Fig. 5). Potential enzyme activities of XYL were significantly decreased with increasing heat application (Fig. 5). No significant alteration by the treatments was observed for AG or NAG (Fig. 5). Overall, potential activities of the seven enzymes showed similar results when assessed at different temperatures (Fig. 5; Fig. S4, Supporting Information). Figure 5. View largeDownload slide Potential enzyme activities (μmol g−1 h−1) of seven different enzymes assayed at 15°C, activation energy (kJ mol−1) of the seven enzymes, and stoichiometry (unit-less) of potential enzyme activities assayed at 15°C for the soils collected from vegetated areas in 2011. The activation energy was calculated using potential enzyme activities assayed at 5°C, 15°C, 25°C and 35°C (Fig. S5, Supporting Information). Results of mixed-effect ANOVAs are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Figure 5. View largeDownload slide Potential enzyme activities (μmol g−1 h−1) of seven different enzymes assayed at 15°C, activation energy (kJ mol−1) of the seven enzymes, and stoichiometry (unit-less) of potential enzyme activities assayed at 15°C for the soils collected from vegetated areas in 2011. The activation energy was calculated using potential enzyme activities assayed at 5°C, 15°C, 25°C and 35°C (Fig. S5, Supporting Information). Results of mixed-effect ANOVAs are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Increasing precipitation reduced activation energy of NAG in all temperature treatments and that of CB only in the control temperature (Fig. 5) supported by a temperature × precipitation effect (P = 0.022). Activation energy of PHOS was higher for the drought and wet treatments than the ambient, as supported by a precipitation effect (P = 0.012, Fig. 5). Increasing precipitation had higher BG:N and C:N ratios, and lower N:P ratios (P = 0.007, 0.018 and 0.034, respectively, Fig. 5). Ratios of BG:P and C:P were lower in the drought and wet treatments than the control (P = 0.007 for both ratios, Fig. 5). A redundancy analysis was conducted to analyze the relative abundances of the major bacterial taxa (Figs 3 and 4) and the activation energy and stoichiometry significantly altered by the precipitation and/or temperature treatments in the 2011 vegetated soils (Fig. 6). Primary sample scores were similar to those in UniFrac (Fig. 2) and dbRDA (Fig. 3), where the scores in the axis 1 co-varied with precipitation quantities (P ≤ 0.001), and increasing temperature decreased the axis 1 scores (P = 0.005). These treatment effects were driven by an increasing relative abundance of Actinobacteria with decreasing precipitation quantities and with warming (Fig. 6). The higher relative abundance of Actinobacteria was associated with increasing NAG activation energy, and N:P and C:P enzyme ratios, and with decreasing CB activation energy and C:N enzyme ratio (Fig. 6). No significant treatment effect was found for the axis 2 scores (Fig. 6). Axes 1 and 2 explained 35.1% of the variability of the data (Fig. 6). Figure 6. View largeDownload slide Results of redundancy analysis using composition of bacterial taxa in the soils collected from vegetated areas in 2011, and the enzyme activation energy and stoichiometry values that were found significantly affected by the experimental treatments in the mixed-effect ANOVAs (Fig. 4). The result of mixed-effect ANOVA for each axis of sample scores is shown in the corresponding panel (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Bactero: Bacteroidetes, Gemmati: Gemmatimonadetes, Proteo; Proteobacteria Figure 6. View largeDownload slide Results of redundancy analysis using composition of bacterial taxa in the soils collected from vegetated areas in 2011, and the enzyme activation energy and stoichiometry values that were found significantly affected by the experimental treatments in the mixed-effect ANOVAs (Fig. 4). The result of mixed-effect ANOVA for each axis of sample scores is shown in the corresponding panel (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Bactero: Bacteroidetes, Gemmati: Gemmatimonadetes, Proteo; Proteobacteria DISCUSSION Manipulations of precipitation and temperature significantly altered the soil bacterial community structure in the vegetated plots, and some biogeochemical properties co-varied with the bacterial structural change. However, the climate manipulations had little impact on the bacterial community structure in bare soils despite significant cumulative treatment effects on soil properties including the C:N ratio. The treatment effects on the fungal community structure were not apparent in either vegetated or bare soils. Our results emphasize that the primary responses of soil bacterial communities to changes in precipitation and temperature were mediated through responses of the vascular plant community, as opposed to being direct responses to the changes in abiotic conditions. Higher relative abundances of Actinobacteria in vegetated soils in less-watered treatments were the dominant change in the bacterial community structure in both 2009 and 2011. The Actinobacteria phylum has many members known to be resistant to drought and able to grow under dry conditions (Goodfellow and Williams 1983; Zvyagintsev et al.2007); some studies have reported higher relative abundances of Actinobacteria in dry soil conditions (Cruz-Martínez et al.2012; Barnard, Osborne and Firestone 2013). Actinobacteria are Gram-positive bacteria, which possess stronger cell walls than Gram-negative bacteria, and are thus considered resistant to desiccation (Schimel, Balser and Wallenstein 2007). Some Actinobacteria members can produce stress-resistant spores (Ait Barka et al.2016), which can help them persist in drought (Santos-Medellin et al.2017; Taketani et al.2017, but see Naylor et al.2017). These characteristics might contribute to the high relative abundances of Actinobacteria in less-watered soils in this study. The co-variation between relative abundances of Actinobacteria and the precipitation quantities in this study appears to have been mediated by the responses of the vascular plant community to the precipitation treatments, given that no such trend was observed in the bare soils in 2011 (Fig. 3). This contrasting precipitation effect between the vegetated and bare soils is consistent with two recent reports by Naylor et al. (2017) and Santos-Medellin et al. (2017); both studies demonstrated that drought significantly increased relative abundances of Actinobacteria in rhizosphere soils, but such trends were limited in bulk soils. One plausible mechanism behind the co-variation was conserved common responses of vascular plant species to drought (Naylor et al.2017). This is supported by two studies showing that increased relative abundances of Actinobacteria in the rhizosphere as well as the root endosphere induced by drought were observed consistently in four rice cultivars (Santos-Medellin et al.2017), and 18 plant species belonging to the Poaceae family in addition to tomato, which was used as an outgoup (Naylor et al.2017). Thus, it is possible that many plant species in the less-watered plots in this study increased production of biochemical compounds in response to water stress, which in turn increased relative Actinobacteria abundance. Another potential mechanism contributing to the consistent co-variation was altered plant community structure. At the onset of the experiment in 2008, the old field experimental site was dominated by C3 perennials (Hoeppner and Dukes 2012). By 2010, non-perennials and C4 plants became more common in the ambient and wet than the drought treatment (Hoeppner and Dukes 2012). Plant species differentially shape rhizosphere properties via root exudates (Hertenberger, Zampach and Bachmann 2002), litter quality and quantity (de Deyn, Cornelissen and Bardgett 2008; Meier and Bowman 2008), nutrient uptakes (Bell et al.2015), mycorrhizal association (Rillig and Mummey 2006), pH (Herr et al.2007) and soil aggregates (Haynes and Beare 1997). The differences in plant community structure among the precipitation and temperature treatments could have influenced rhizosphere properties, which in turn resulted in differences in the microbial community structure (el Zahar Haichar et al.2008; Philippot et al.2013; Tkacz et al.2015). Nitrogen dynamics under vegetation could also be a driver for the observed differences in the bacterial community structure. Inorganic N amendment to mineral soils diverse in origin consistently increased relative abundances of Actinobacteria (Ramirez, Craine and Fierer 2012). This result was supported by a greenhouse experiment conducted by Bell et al. (2015) who demonstrated that relative abundances of Actinobacteria consistently declined during the growing season as available N in soils decreased due to N assimilation by four different grass species. In our study, indeed, inorganic N concentrations in vegetated soils of 2011 were higher in the drought than ambient or wet precipitation treatments (Fig. S5, Supporting Information), and relative abundances of Actinobacteria were significantly correlated with inorganic N concentration (Fig. S6, Supporting Information). Relative quantities of available C altered by the response of plant community to the precipitation treatments might not be a contributor to the observed microbial community shifts, for two reasons. First, increased available C should, in theory, favor copiotrophs over oligotrophs, but the phylum Actinobacteria may not be classified as either of the categories (Fierer, Bradford and Jackson 2007). Second, net primary productivity, which should affect available C in the vegetated soils, was not significantly altered by the precipitation treatments (Hoeppner and Dukes 2012). Differences in available C should be found, instead, between the vegetated and bare soils; we hypothesized that copiotrophs should be favored in the vegetated soils that were supplied with C via rhizodeposition from the vegetation, and oligotrophs were favored in bare soils that had less available C without organic matter supply from vascular plants. Our results supported this hypothesis. The vegetated and bare soils collected in 2011 had higher relative abundances of β-Proteobacteria and Acidobacteria, respectively (Fig. 3), which are considered as copiotrophic and oligotrophic taxa (Fierer, Bradford and Jackson 2007). The constituents of rhizodeposition, such as root exudates, are C-rich low molecular-weight compounds released by living roots (Hütsch, Augustin and Merbach 2002), and are immediately assimilated as a C source by rhizosphere microbes upon production (Denef et al.2009), and thus should support copiotrophs over oligotrophs. We do not know the mechanisms behind the higher relative abundances of α-Proteobacteria in the bare than the vegetated soils (Fig. 2). This observation contrasts with the higher α-Proteobacteria in vegetated than bare soils reported by Thomson et al. (2010) and Yergeau et al. (2012). It is possible that relative abundance of α-Proteobacteria was passively increased in the bare soils due to the increased relative abundances of copiotrophs such as β-Proteobacteria and Bacteroidetes (Fierer, Bradford and Jackson 2007) under vegetation. Mycorrhizal fungi associated with the plant community could interact with the rhizosphere bacterial community (Bonfante and Anca 2009). Nuccio et al. (2013) reported that the presence of arbuscular mycorrhizal fungi (phylum Glomeromycota) reduced relative abundances of Actinobacteria. Our data of relative abundance of Glomeromycota show a trend consistent with potential influence of Glomeromycota on the Actinobacteria population in the high-temperature treatment, but not in the control (Fig. S7, Supporting Information). It is possible that absolute abundances of Glomeromycota were positively correlated with quantities of precipitation, which in turn affected the relative abundance of Actinobacteria. The significant temperature effect on bacterial community structure in the 2011 soils was also mediated by the presence of plants (Figs 2 and 3). The lack of temperature effects on bacterial community structure in the bare soils is consistent with results of Oliverio, Bradford and Fierer (2017) who investigated bacterial response to increased temperature in bare soils. In the study, a group of bacterial OTUs consistently responded to increased temperature in a lab incubation in bare soils collected from diverse ecosystems. However, the OTUs did not appear to be phylogenetically predictable, consistent with our results of no significant response to increased temperature in the bare soils using UniFrac (Fig. 2) or dbRDA (Fig. 3). Thus, it appears that the soil microbial community structure is not directly sensitive to ecosystem warming. Five of the seven potential enzyme activities assessed were significantly affected by the precipitation and/or temperature treatments, where BG, CB, LAP and PHOS showed a similar trend; the potential activities were lower in plots with lower precipitation and higher temperature (Fig. 5). One of the primary factors affecting potential enzyme activities is population sizes of soil microbes that produce extracellular enzymes (Nannipieri, Kandeler and Ruggiero 2002); thus microbial biomass can be positively correlated with potential enzyme activities (e.g. Waldrop, McColl and Powers 2003; Tan, Chang and Kabzems 2008; Keeler, Hobbie and Kellog 2009; Brockett, Prescott and Grayston 2012, but see Bell, Klironomos and Henry 2010). The four potential enzyme activities were, indeed, positively correlated with substrate induced respiration (Fig. S8, Supporting Information), an index of microbial biomass (Anderson and Domsch 1978). On the other hand, properties of the potential enzyme activities, including temperature sensitivity and stoichiometry, should reflect microbial community composition, which controls enzyme qualities (Alster et al. 2016a, b). We found that four enzyme properties co-varied with the primary axis reflecting microbial community structure (Fig. 6). Though the co-variations are only correlations, they indicate that differences in bacterial community structure could affect the extracellular enzyme properties in the vegetated soils. This result makes this one of the few studies to link microbial community structure and soil extracellular enzyme activities (Bier et al.2015). We do not know the mechanisms behind our observation that the fungal community structure did not respond to the precipitation or temperature treatments with the same magnitude as the bacterial community (Fig. 2). Reponses of bacterial and fungal community structure to environmental manipulations, such as precipitation and temperature, can depend on the type of ecosystem and manipulation, as various results have been reported. In responses to precipitation manipulations, for instance, some studies showed trends similar to this study with a significant responses of community structure for bacteria, but to lesser degrees for fungi (Barnard, Osborne and Firestone 2013; Amend et al.2016; Zhang et al.2016; Naylor et al.2017). However, some studies reported different trends, including an opposite trend with little effect on bacteria but stronger effects on fungi (Castro et al.2010), bacteria and fungi both being significantly affected (McHugh and Schwartz 2015; Waring and Hawkes 2015; Santos-Medellin et al.2017), and neither of them being altered (McHugh and Schwartz 2015). In addition to precipitation quantities, timing of precipitation manipulations can be an important factor; in a Mediterranean-type grassland, increased precipitation during rainy seasons significantly altered a plant community compared to controls (Suttle, Thomsen and Power 2007), which was accompanied with significant changes in the fungal community structure in soils (Hawkes et al.2011) but not in the bacterial community structure (Cruz-Martí et al.2009). In contrast, in a different Mediterranean-type grassland, the bacterial community structure was significantly changed, but fungal community was not in response to water addition during the dry season (Barnard, Osborne and Firestone 2013). These contrasting responses of bacteria vs. fungi to precipitation manipulations make it a challenge to generalize responses of the two groups to environmental changes (Griffiths and Philippot 2013). Another potential explanation for the lack of fungal community response to the treatments in this study could be the fungal 18S primer set employed in this study, as primer selectivity can cause quantitative biases in amplicons (Peršoh 2015). However, Barnard et al. (2013) and Naylor et al. (2017) used different sets of fungal-specific primers, targeting the 28S and ITS2 regions, respectively, and found little drought effect on a rhizosphere fungal community, whereas responses of the co-occurring rhizosphere bacterial community were similar to those of this study. In conclusion, our results demonstrated that soil microbial communities, which play a major role in C and nutrient biogeochemistry, respond to environmental changes primarily through shifts in plant-soil-microbial interactions, as opposed to responding directly to the changes in abiotic conditions. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. Acknowledgements We thank Drs Guy Beresford and Ann M. Hess for consultation in molecular work and statistics, respectively. We thank Carol Goranson for help at the BACE site. FUNDING This work was supported by grants from the Office of Biological and Environmental Research (BER) of the U.S. Department of Energy's Office of Science to M.D.W. and J.S.D. through the Northeastern Regional Center of the National Institute for Climate Change Research, and from the National Science Foundation (Division of Environmental Biology 0546670 to J.S.D.). Additional support for J.S.D.'s participation in this project was provided by Hatch project 1000026 of the United States Department of Agriculture's National Institute of Food and Agriculture. Conflicts of interest. None declared. REFERENCES Ait Barka E, Vatsa P, Sanchez L et al.   Taxonomy, physiology, and natural products of Actinobacteria. Microbiol Mol Biol Rev  2016; 80: 1– 43. Google Scholar CrossRef Search ADS PubMed  Alster CJ, Baas P, Wallenstein MD et al.   Temperature sensitivity as a microbial trait using parameters from macromolecular rate theory. FrontMicrobiol  2016a; 7: 1– 10. Alster CJ, Koyama A, Johnson NG et al.   Temperature sensitivity of soil microbial communities: an application of macromolecular rate theory to microbial respiration. J Geophys Res:Biogeosci  2016b; 1420– 1433. Amend AS, Martiny AC, Allison SD et al.   Microbial response to simulated global change is phylogenetically conserved and linked with functional potential. ISME J  2016; 10: 109– 18. Google Scholar CrossRef Search ADS PubMed  Anderson JPE, Domsch KH. A physiological method for the quantitative measurement of microbial biomass in soils. Soil Biol Biochem  1978; 10: 215– 21. Google Scholar CrossRef Search ADS   Arrhenius S. Über die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Säuren. Z Phys Chem  1889; 4: 226– 48. Averill C, Turner BL, Finzi AC. Mycorrhiza-mediated competition between plants and decomposers drives soil carbon storage. Nature  2014; 505: 543– 5. Google Scholar CrossRef Search ADS PubMed  Barnard RL, Osborne CA, Firestone MK. Responses of soil bacterial and fungal communities to extreme desiccation and rewetting. ISME J  2013; 7: 2229– 41. Google Scholar CrossRef Search ADS PubMed  Bates ST, Berg-Lyons D, Caporaso JG et al.   Examining the global distribution of dominant archaeal populations in soil. ISME J  2010; 5: 908– 17. Google Scholar CrossRef Search ADS PubMed  Bell CW, Asao S, Calderon F et al.   Plant nitrogen uptake drives rhizosphere bacterial community assembly during plant growth. Soil Biol Biochem  2015; 85: 170– 82. Google Scholar CrossRef Search ADS   Bell TH, Klironomos JN, Henry HA. Seasonal responses of extracellular enzyme activity and microbial biomass to warming and nitrogen addition. Soil Sci Soc Am J  2010; 74: 820– 8. Google Scholar CrossRef Search ADS   Bier RL, Bernhardt ES, Boot CM et al.   Linking microbial community structure and microbial processes: an empirical and conceptual overview. FEMS Microbiol Ecol  2015; 91: fiv113. Google Scholar CrossRef Search ADS PubMed  Blagodatskaya E, Blagodatsky S, Khomyakov N et al.   Temperature sensitivity and enzymatic mechanisms of soil organic matter decomposition along an altitudinal gradient on Mount Kilimanjaro. Sci Rep  2016; 6: 1– 11. Google Scholar CrossRef Search ADS PubMed  Bonfante P, Anca I-A. Plants, mycorrhizal fungi, and bacteria: a network of interactions. Annu Rev Microbiol  2009; 63: 363– 83. Google Scholar CrossRef Search ADS PubMed  Borneman J, Hartin RJ. PCR primers that amplify fungal rRNA genes from environmental samples. Appl Environ Microbiol  2000; 66: 4356– 4360. Google Scholar CrossRef Search ADS PubMed  Brockett BFT, Prescott CE, Grayston SJ. Soil moisture is the major factor influencing microbial community structure and enzyme activities across seven biogeoclimatic zones in western Canada. Soil Biol Biochem  2012; 44: 9– 20. Google Scholar CrossRef Search ADS   Cai W, Borlace S, Lengaigne M et al.   Increasing frequency of extreme El Nino events due to greenhouse warming. Nature Clim Change  2014; 4: 111– 6. Google Scholar CrossRef Search ADS   Campbell CD, Chapman SJ, Cameron CM et al.   A rapid microtiter plate method to measure carbon dioxide evolved from carbon substrate amendments so as to determine the physiological profiles of soil microbial communities by using whole soil. Appl Environ Microbiol  2003; 69: 3593– 9. Google Scholar CrossRef Search ADS PubMed  Caporaso JG, Bittinger K, Bushman FD et al.   PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics  2010b; 26: 266– 7. Google Scholar CrossRef Search ADS   Caporaso JG, Kuczynski J, Stombaugh J et al.   QIIME allows analysis of high-throughput community sequencing data. Nat Methods  2010a; 7: 335– 6. Google Scholar CrossRef Search ADS   Castro HF, Classen AT, Austin EE et al.   Soil microbial community responses to multiple experimental climate change drivers. Appl Environ Microb  2010; 76: 999– 1007. Google Scholar CrossRef Search ADS   Cole JR, Wang Q, Cardenas E et al.   The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res  2009; 37: D141– 5. Google Scholar CrossRef Search ADS PubMed  Chen I-C, Hill JK, Ohlemüller R et al.   Rapid range shifts of species associated with high levels of climate warming. Science  2011; 333: 1024– 6. Google Scholar CrossRef Search ADS PubMed  Cotrufo MF, Soong JL, Horton AJ et al.   Formation of soil organic matter via biochemical and physical pathways of litter mass loss. Nat Geosci  2015; 8: 776. Google Scholar CrossRef Search ADS   Cowles JM, Wragg PD, Wright AJ et al.   Shifting grassland plant community structure drives positive interactive effects of warming and diversity on aboveground net primary productivity. Glob Change Biol  2016; 22: 741– 9. Google Scholar CrossRef Search ADS   Cruz-Martínez K, Rosling A, Zhang Y et al.   Effect of rainfall-induced soil geochemistry dynamics on grassland soil microbial communities. Appl Environ Microb  2012; 78: 7587– 95. Google Scholar CrossRef Search ADS   Cruz-Martinez K, Suttle KB, Brodie EL et al.   Despite strong seasonal responses, soil microbial consortia are more resilient to long-term changes in rainfall than overlying grassland. ISME J  2009; 3: 738– 44. Google Scholar CrossRef Search ADS PubMed  de Deyn GB, Cornelissen JHC, Bardgett RD. Plant functional traits and soil carbon sequestration in contrasting biomes. Ecol Lett  2008; 11: 516– 31. Google Scholar CrossRef Search ADS PubMed  Denef K, Roobroeck D, Manimel Wadu MCW et al.   Microbial community composition and rhizodeposit-carbon assimilation in differently managed temperate grassland soils. Soil Biol Biochem  2009; 41: 144– 53. Google Scholar CrossRef Search ADS   el Zahar Haichar F, Marol C, Berge O et al.   Plant host habitat and root exudates shape soil bacterial community structure. ISME J  2008; 2: 1221– 30. Google Scholar CrossRef Search ADS PubMed  Evans SE, Wallenstein MD. Soil microbial community response to drying and rewetting stress: does historical precipitation regime matter? Biogeochemistry  2012; 109: 101– 16. Google Scholar CrossRef Search ADS   Falkowski PG, Fenchel T, Delong EF. The microbial engines that drive Earth's biogeochemical cycles. Science  2008; 320: 1034– 9. Google Scholar CrossRef Search ADS PubMed  Feng X, Simpson MJ. Temperature and substrate controls on microbial phospholipid fatty acid composition during incubation of grassland soils contrasting in organic matter quality. Soil Biol Biochem  2009; 41: 804– 12. Google Scholar CrossRef Search ADS   Fierer N, Bradford MA, Jackson RB. Toward an ecological classification of soil bacteria. Ecology  2007; 88: 1354– 64. Google Scholar CrossRef Search ADS PubMed  Fierer N, Jackson RB. The diversity and biogeography of soil bacterial communities. P Natl Acad Sci USA  2006; 103: 626– 31. Google Scholar CrossRef Search ADS   Gibbons SM, Lekberg Y, Mummey DL et al.   Invasive plants rapidly reshape soil properties in a grassland ecosystem. mSystems  2017; 2: 1– 13. Google Scholar CrossRef Search ADS   Goodfellow M, Williams ST. Ecology of Actinomycetes. Annu Rev Microbiol  1983; 37: 189– 216. Google Scholar CrossRef Search ADS PubMed  Griffiths BS, Philippot L. Insights into the resistance and resilience of the soil microbial community. FEMS Microbiol Rev  2013; 37: 112– 29. Google Scholar CrossRef Search ADS PubMed  Hamady M, Lozupone C, Knight R. Fast UniFrac: facilitating high-throughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J  2010; 4: 17– 27. Google Scholar CrossRef Search ADS PubMed  Haynes R, Beare M. Influence of six crop species on aggregate stability and some labile organic matter fractions. Soil Biol Biochem  1997; 29: 1647– 53. Google Scholar CrossRef Search ADS   Hawkes CV, Kivlin SN, Rocca JD et al.   Fungal community responses to precipitation. Glob Change Biol  2011; 17: 1637– 45. Google Scholar CrossRef Search ADS   Herr C, Chapuis, LL, Dassonville N et al.   Seasonal effect of the exotic invasive plant Solidago gigantea on soil pH and P fractions. J Plant Nutr Soil Sci  2007; 170: 729– 38. Google Scholar CrossRef Search ADS   Hertenberger G, Zampach P, Bachmann G. Plant species affect the concentration of free sugars and free amino acids in different types of soil. J Plant Nutr Soil Sci  2002; 165: 557– 65. Google Scholar CrossRef Search ADS   Hoeppner SS, Dukes JS. Interactive responses of old-field plant growth and composition to warming and precipitation. Global Change Biol  2012; 18: 1754– 68. Google Scholar CrossRef Search ADS   Hütsch BW, Augustin J, Merbach W. Plant rhizodeposition —an important source for carbon turnover in soils. J Plant Nutr Soil Sci  2002; 165: 397– 407. Google Scholar CrossRef Search ADS   Johnson M, Zaretskaya I, Raytselis Y et al.   NCBI BLAST: a better web interface. Nucleic Acids Research  2008; 36: W5– W9. Google Scholar CrossRef Search ADS PubMed  Judd CR, Koyama A, Simmons MP et al.   Co-variation in methanotroph community composition and activity in three temperate grassland soils. Soil Biol Biochem  2016; 95: 78– 86. Google Scholar CrossRef Search ADS   Kardol P, Campany CE, Souza L et al.   Climate change effects on plant biomass alter dominance patterns and community evenness in an experimental old‐field ecosystem. Global Change Biol  2010; 16: 2676– 87. Google Scholar CrossRef Search ADS   Keeler B, Hobbie S, Kellogg L. Effects of long-term nitrogen addition on microbial enzyme activity in eight forested and grassland sites: implications for litter and soil organic matter decomposition. Ecosystems  2009; 12: 1– 15. Google Scholar CrossRef Search ADS   Kelly AE, Goulden ML. Rapid shifts in plant distribution with recent climate change. P Natl Acad Sci  2008; 105: 11823– 6. Google Scholar CrossRef Search ADS   Koyama A, Wallenstein MD, Simpson RT et al.   Carbon-degrading enzyme activities stimulated by increased nutrient availability in arctic tundra soils. PLoS One  2013; 8: e77212. Google Scholar CrossRef Search ADS PubMed  Koyama A, Wallenstein MD, Simpson RT et al.   Soil bacterial community composition altered by increased nutrient availability in Arctic tundra soils. Front Microbiol  2014; 5: 1– 16. Google Scholar CrossRef Search ADS PubMed  Lauber CL, Hamady M, Knight R et al.   Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl Environ Microb  2009; 75: 5111– 20. Google Scholar CrossRef Search ADS   Legay N, Baxendale C, Grigulis K et al.   Contribution of above-and below-ground plant traits to the structure and function of grassland soil microbial communities. Ann Bot  2014; 114: 1011– 21. Google Scholar CrossRef Search ADS PubMed  Legendre P, Anderson MJ. Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecol Monogr  1999; 69: 1– 24. Google Scholar CrossRef Search ADS   Levine UY, Teal TK, Robertson GP et al.   Agriculture's impact on microbial diversity and associated fluxes of carbon dioxide and methane. ISME J  2011; 5: 1683– 91. Google Scholar CrossRef Search ADS PubMed  Louca S, Jacques SMS, Pires APF et al.   High taxonomic variability despite stable functional structure across microbial communities. Nat Ecol Evol  2017; 1: 0015. Google Scholar CrossRef Search ADS   Makino W, Cotner J, Sterner R et al.   Are bacteria more like plants or animals? Growth rate and resource dependence of bacterial C: N: P stoichiometry. Funct Ecol  2003; 17: 121– 30. Google Scholar CrossRef Search ADS   McHugh TA, Schwartz E. Changes in plant community composition and reduced precipitation have limited effects on the structure of soil bacterial and fungal communities present in a semiarid grassland. Plant Soil  2015; 388: 175– 86. Google Scholar CrossRef Search ADS   Meier CL, Bowman WD. Links between plant litter chemistry, species diversity, and below-ground ecosystem function. P Natl Acad Sci  2008; 105: 19780– 5. Google Scholar CrossRef Search ADS   Min S-K, Zhang X, Zwiers FW et al.   Human contribution to more-intense precipitation extremes. Nature  2011; 470: 378– 81. Google Scholar CrossRef Search ADS PubMed  Moore JC, Berlow EL, Coleman DC et al.   Detritus, trophic dynamics and biodiversity. Ecol Lett  2004; 7: 584– 600. Google Scholar CrossRef Search ADS   Mueller K, Blumenthal DM, Pendall E et al.   Impacts of warming and elevated CO2 on a semi‐arid grassland are non‐additive, shift with precipitation, and reverse over time. Ecol Lett  2016; 19: 956– 66. Google Scholar CrossRef Search ADS PubMed  Naylor D, DeGraaf S, Purdom E et al.   Drought and host selection influence bacterial community dynamics in the grass root microbiome. ISME J  2017; 11: 2691– 704. Google Scholar CrossRef Search ADS PubMed  Nuccio EE, Hodge A, Pett-Ridge J et al.   An arbuscular mycorrhizal fungus significantly modifies the soil bacterial community and nitrogen cycling during litter decomposition. Environ Microbiol  2013; 15: 1870– 81. Google Scholar CrossRef Search ADS PubMed  Nannipieri P, Kandeler E, Ruggiero P. Enzyme activities and microbiological and biochemical processes in soil. In Enzymes in the Environment , New York: Marcel Dekker, 2002, 1– 33. Oliverio AM, Bradford MA, Fierer N. Identifying the microbial taxa that consistently respond to soil warming across time and space. Global Change Biol  2017; 23: 2117– 29. Google Scholar CrossRef Search ADS   Peršoh D. Plant-associated fungal communities in the light of meta’omics. Fungal Divers  2015; 75: 1– 25. Google Scholar CrossRef Search ADS   Philippot L, Raaijmakers JM, Lemanceau P et al.   Going back to the roots: the microbial ecology of the rhizosphere. Nat Rev Micro  2013; 11: 789– 99. Google Scholar CrossRef Search ADS   Price MN, Dehal PS, Arkin AP. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol Biol Evol  2009; 26: 1641– 50. Google Scholar CrossRef Search ADS PubMed  R Core Team, R: A Language and Environment for Statistical Computing . Vienna, Austria: R Foundation for Statistical Computing, 2015, URL https://www.R-project.org/. Ramirez KS, Craine JM, Fierer N. Consistent effects of nitrogen amendments on soil microbial communities and processes across biomes. Global Change Biol  2012; 18: 1918– 27. Google Scholar CrossRef Search ADS   Rillig MC, Mummey DL. Mycorrhizas and soil structure. New Phytol  2006; 171: 41– 53. Google Scholar CrossRef Search ADS PubMed  Romaní AM, Fischer H, Mille-Lindblom C et al.   Interactions of bacteria and fungi on decomposing litter: differential extracellular enzyme activities. Ecology  2006; 87: 2559– 69. Google Scholar CrossRef Search ADS PubMed  Rousk J, Bååth E. Growth of saprotrophic fungi and bacteria in soil. FEMS Microbiol Ecol  2011; 78: 17– 30. Google Scholar CrossRef Search ADS PubMed  Rousk J, Bååth E, Brookes PC et al.   Soil bacterial and fungal communities across a pH gradient in an arable soil. ISME J  2010; 4: 1340– 51. Google Scholar CrossRef Search ADS PubMed  Rudisill MA, Turco RF, Hoagland LA. Fertility practices and rhizosphere effects alter ammonia oxidizer community structure and potential nitrification activity in pepper production soils. Appl Soil Ecol y  2016; 99: 70– 7. Google Scholar CrossRef Search ADS   Saiya-Cork KR, Sinsabaugh RL, Zak DR. The effects of long term nitrogen deposition on extracellular enzyme activity in an Acer saccharum forest soil. Soil Biol Biochem  2002; 34: 1309– 15. Google Scholar CrossRef Search ADS   Santos-Medellín C, Edwards J, Liechty Z et al.   Drought stress results in a compartment-specific restructuring of the rice root-associated microbiomes. mBio  2017; 8: 1– 15. Google Scholar CrossRef Search ADS   Schimel J, Balser TC, Wallenstein M. Microbial stress‐response physiology and its implications for ecosystem function. Ecology  2007; 88: 1386– 94. Google Scholar CrossRef Search ADS PubMed  Six J, Frey SD, Thiet RK et al.   Bacterial and fungal contributions to carbon sequestration in agroecosystems. Soil Sci Soc Am J  2006; 70: 555– 69. Google Scholar CrossRef Search ADS   Steinweg JM, Dukes JS, Wallenstein MD. Modeling the effects of temperature and moisture on soil enzyme activity: Linking laboratory assays to continuous field data. Soil Biology and Biochemistry  2012; 55: 85– 92. Google Scholar CrossRef Search ADS   Steinweg JM, Dukes JS, Paul EA et al.   Microbial responses to multi-factor climate change: effects on soil enzymes. Front Microbiol  2013; 4: 1– 11. Google Scholar CrossRef Search ADS PubMed  Suseela V, Conant RT, Wallenstein MD et al.   Effects of soil moisture on the temperature sensitivity of heterotrophic respiration vary seasonally in an old-field climate change experiment. Global Change Biol  2012; 18: 336– 48. Google Scholar CrossRef Search ADS   Suseela V, Tharayil N, Xing B et al.   Labile compounds in plant litter reduce the sensitivity of decomposition to warming and altered precipitation. New Phytol  2013; 200: 122– 33. Google Scholar CrossRef Search ADS PubMed  Taketani RG, Lançoni MD, Kavamura VN et al.   Dry Season Constrains Bacterial Phylogenetic Diversity in a Semi-Arid Rhizosphere System. Microbial Ecology  2017; 73: 153– 61. Google Scholar CrossRef Search ADS PubMed  Tan X, Chang SX, Kabzems R. Soil compaction and forest floor removal reduced microbial biomass and enzyme activities in a boreal aspen forest soil. Biol Fert Soils  2008; 44: 471– 9. Google Scholar CrossRef Search ADS   Thomson BC, Ostle N, McNamara N et al.   Vegetation affects the relative abundances of dominant soil bacterial taxa and soil respiration rates in an upland grassland soil. Microbial Ecol  2010; 59: 335– 43. Google Scholar CrossRef Search ADS   Tkacz A, Cheema J, Chandra G et al.   Stability and succession of the rhizosphere microbiota depends upon plant type and soil composition. ISME J  2015; 9: 2349– 59. Google Scholar CrossRef Search ADS PubMed  Suttle K, Thomsen MA, Power ME. Species interactions reverse grassland responses to changing climate. Science  2007; 315: 640– 2. Google Scholar CrossRef Search ADS PubMed  Waldrop M, McColl J, Powers R. Effects of forest postharvest management practices on enzyme activities in decomposing litter. Soil Sci Soc Am J  2003; 67: 1250– 6. Google Scholar CrossRef Search ADS   Wallenstein MD, Weintraub MN. Emerging tools for measuring and modeling the in situ activity of soil extracellular enzymes. Soil Biol Biochem  2008; 40: 2098– 106. Google Scholar CrossRef Search ADS   Waring BG, Hawkes CV. Short-term precipitation exclusion alters microbial responses to soil moisture in a wet tropical forest. Microbial Ecol  2015; 69: 843– 54. Google Scholar CrossRef Search ADS   Yang H, Li Y, Wu M et al.   Plant community responses to nitrogen addition and increased precipitation: the importance of water availability and species traits. Global Change Biol  2011; 17: 2936– 44. Google Scholar CrossRef Search ADS   Yergeau E, Bokhorst S, Kang S et al.   Shifts in soil microorganisms in response to warming are consistent across a range of Antarctic environments. ISME J  2012; 6: 692– 702. Google Scholar CrossRef Search ADS PubMed  Zhang K, Shi Y, Jing X et al.   Effects of short-term warming and altered precipitation on soil microbial communities in alpine grassland of the Tibetan plateau. Front Microbiol  2016; 7: 1– 11. Google Scholar PubMed  Zvyagintsev D, Zenova G, Doroshenko E et al.   Actinomycete growth in conditions of low moisture. Biol Bull  2007; 34: 242– 7. Google Scholar CrossRef Search ADS   © FEMS 2017. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png FEMS Microbiology Ecology Oxford University Press

Soil bacterial community responses to altered precipitation and temperature regimes in an old field grassland are mediated by plants

Loading next page...
 
/lp/ou_press/soil-bacterial-community-responses-to-altered-precipitation-and-NUClXwCfde
Publisher
Oxford University Press
Copyright
© FEMS 2017. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0168-6496
eISSN
1574-6941
D.O.I.
10.1093/femsec/fix156
Publisher site
See Article on Publisher Site

Abstract

Abstract The structure and function of soil microbiomes often change in response to experimental climate manipulations, suggesting an important role in ecosystem feedbacks. However, it is difficult to know if microbes are responding directly to environmental changes or are more strongly impacted by plant responses. We investigated soil microbial responses to precipitation and temperature manipulations at the Boston-Area Climate Experiment in Massachusetts, USA, in both vegetated and bare plots to parse direct vs. plant-mediated responses to multi-factor climate change. We assessed the bacterial community in vegetated soils in 2009, two years after the experiment was initiated, and bacterial and fungal community in vegetated and bare soils in 2011. The bacterial community structure was significantly changed by the treatments in vegetated soils. However, such changes in the bacterial community across the treatments were absent in the 2011 bare soils. These results suggest that the bacterial communities in vegetated soils were structured via plant community shifts in response to the abiotic manipulations. Co-variation between bacterial community structure and temperature sensitivities and stoichiometry of potential enzyme activities in the 2011 vegetated soils suggested a link between bacterial community structure and ecosystem function. This study emphasizes the importance of plant-soil-microbial interactions in mediating responses to future climate change. Boston-Area Climate Experiment, precipitation, temperature, bacterial and fungal community structure, potential enzyme activity INTRODUCTION Soil microbial communities appear to be sensitive to climatic changes that are characterized by increasing temperature and intensified precipitation regimes in many parts of the world (Min et al.2011; Cai et al.2014). However, it is unclear whether responses of microbiome structure and function in experimental climate change are direct or are mediated by simultaneous plant community shifts in response to the abiotic changes (Legay et al.2014). It is well established that abiotic factors such as pH (Fierer and Jackson 2006; Lauber et al.2009), temperature (Feng and Simpson 2009) and moisture availability (Brockett, Prescott and Grayston 2012) are major drivers of microbial community structure. Plant community shifts in response to climate change factors have been observed in natural systems (Kelly and Goulden 2008; Chen et al.2011) and in experimental manipulations of precipitation (Kardol et al.2010; Yang et al.2011) and temperature (Cowles et al.2016; Mueller et al.2016). Such shifts in vascular plant communities can shape the microbial community structure in the rhizosphere (Philippot et al.2013). Therefore, both abiotic factors and plant community shifts could structure soil microbial communities under changing climates. Changes in microbial community structure in response to the environmental changes could also alter biogeochemical functioning. Soil microbes, including bacteria and fungi, play vital roles in biogeochemical cycling of carbon (C) and nutrients such as nitrogen (N) and phosphorus (P) (Falkowski, Fenchel and Delong 2008). These roles include serving as the base of the soil food web (Moore et al.2004), helping form and stabilize soil organic matter (Averill, Turner and Finzi 2014; Cotrufo et al.2015), and mineralizing of soil organic matter, which makes available nutrients for plants and microbes themselves (Romaní et al.2006). As providers of the initial, rate-limiting step of organic matter mineralization, soil microbes produce extracellular enzymes to depolymerize and solubilize large organic compounds (Wallenstein and Weintraub 2008). Characteristics of extracellular enzymes, such as their temperature sensitivity (Alster et al.2016a), and stoichiometry among C and nutrients (Makino et al.2003), vary among different soil microbes. Therefore, shifts in soil microbial communities potentially affect soil enzymatic properties (Blagodatskaya et al.2016), even though it is possible that significantly different microbial communities can result in similar functions (Gibbons et al. 2017; Louca et al.2017). Although co-variation between microbial community and biogeochemical processes, such as nitrifiers and nitrification (e.g. Rudisill, Turco and Hoagland 2016) and methanotrophs and methane oxidation (e.g. Levine et al.2011, Judd et al.2016), has been reported, such relationships between microbial community structure and soil enzyme properties have not been well explored (Bier et al.2015), In this study, we investigated how manipulations of precipitation and temperature influenced microbial community structure and biogeochemical properties in soils at the Boston-Area Climate Experiment (BACE) in a mesic old-field community in Massachusetts, USA. We assessed the community structure of bacteria in vegetated soils in June 2009, two years after the experiment was initiated, and that of bacteria and fungi in both vegetated and bare soils in July 2011 during the growing season, using Roche 454 sequencing. We hypothesized that (i) microbial populations in the vegetated soils were more affected by the precipitation and temperature treatments than those in bare soils, and (ii) bacterial and fungal communities responded to the treatments differently as the two taxa possess distinctive physiological differences such as biomass turnover rates (Rousk and Bååth 2011) and carbon use efficiency (Six et al.2006). We measured soil properties including water content, organic C and total N, and substrate-induced respiration as an index for microbial biomass. To investigate potential effects of microbial community structure on biogeochemical processes, we quantified activities of seven extracellular enzymes involved in hydrolysis of C and N compounds and phosphate, in vegetated soils collected in 2011. The enzyme assay was conducted at four different temperatures in the lab to assess temperature sensitivities of each of the seven enzymes. We explored the relationship between microbial community structure and enzyme stoichiometry and temperature sensitivity via a multivariate analysis. MATERIALS AND METHODS Study site The study was conducted at the BACE site, located in an old-field community in Waltham, Massachusetts, USA (42°23΄ 3″N, 71° 12΄ 52″W). Mean annual temperature and precipitation are 9.5°C and 1194 mm, respectively (Hoeppner and Dukes 2012). Three levels of precipitation and four levels of temperature treatments had been applied to 2 m by 2 m plots in a randomized, complete block, split-plot design (N = 3) since 2007 and 2008, respectively. This experimental design resulted in a total of 36 plots, which were the study and statistical units. The three precipitation treatments were drought (50% of ambient precipitation, year-round), ambient, and wet (150% of ambient during the growing season). The four temperature treatments were control, low (+∼1°C), medium (+∼2.7°C) and high (+∼4°C) warming using ceramic infrared heaters (Hoeppner and Dukes 2012). The study site had a loamy top soil (Suseela et al.2013). At the onset of the experiment in 2007, the vegetation was a mixture of native and introduced forbs and grasses, where C3 species, perennials and grasses were dominant over C4 species, annuals, and forbs, respectively (Hoeppner and Dukes 2012). By the growing season of 2010, the precipitation and temperature treatments had affected the plant community, where C4 grasses and other annual and biennial species entered the C3 perennial-dominated community in the ambient and wet treatments and the warmed treatments (Hoeppner and Dukes 2012). However, the precipitation and warming treatments did not significantly alter total plant production (Hoeppner and Dukes 2012). Soil sampling, processing and analyses Surface soils to 5 cm in depth were collected from vegetated areas in June 2009, and, in July 2011, from vegetated areas as well as bare surfaces within PVC collars (25 and 30 cm in diameter and depth, respectively) that excluded growth of vascular plants (Suseela et al.2012), in all 36 plots. The soil samples were transported on ice overnight to Natural Resource Ecology Laboratory, Colorado State University, Colorado, USA. The soils were sieved with a 2 mm screen and kept at 4°C for subsequent measurements. Each soil sample was quantified for soil water content (SWC), soil organic carbon (SOC), total nitrogen (N) and substrate-induced respiration rate. SWCs were measured by drying the soils at 105°C for 48 h. For SOC and total N analyses, soils were first dried out at 60°C, and ground using a Brinkmann Retsch mill (Haan, Germany), followed by quantification with a LECO TruSpec® (Leco Corporation, St. Joseph, Michigan, USA). Substrate-induced respiration rates were measured using MicroRespTM (Aberdeen, UK, Campbell et al.2003) in the same manner described in detail by Steinweg et al. (2013). Briefly, each sieved soil was amended with deionized water to 55% of the water holding capacity, and added to a deep 96-well plate with three technical replicates. Following incubation at 4°C for 18 h, 25 μL of 1 M glucose solution was added to each well and incubated at 15°C for 6 h. The same volume of deionized water was added to each soil in another plate. Differences in quantities of CO2 produced between the glucose- and deionized water-amended sub-samples during the 6-h incubation were calculated as substrate induced respiration. Enzyme assays Potential enzyme activities were assessed for the vegetated soils collected in 2011 using a fluorometric method (Saiya-Cork, Sinsabaugh and Zak 2002) modified by Steinweg et al. (2012) in the same manner described in details by Koyama et al. (2013). In total, seven enzymes were assessed: four enzymes to hydrolyze C-rich substrates, including β-glucosidase (BG), cellobiohydrolase (CB), xylosidase (XYL) and α-glucosidase (AG), two enzymes to hydrolyze N-rich substrates including n-acetyl-glucosaminidase (NAG) and leucine-amino-peptidase (LAP), and phosphatase (PHOS). Briefly, each soil sub-sample of 2.75 g was homogenized in 91 mL of 50 mM sodium acetate (pH 5.5) in a blender. An 800 μL aliquot of the slurry was pipetted into each of seven wells in a column of a 96 deep-well plate, and mixed with 200 μL solution of each of the seven substrates. Substrate concentrations had been pre-determined so that they were not completely consumed during incubation. Reference standards were prepared in a similar manner; 200 μL of fluorescent standard, ranging from zero to 200 μM, were mixed with an aliquot of 800 μL slurry for each soil. Two types of standards were used; 7-amino-4-methylcoumarin (MUC) for LAP and 4-methylumbelliferone (MUB) for the other enzymes. Four plates were prepared for each soil sample to assess potential enzyme activities at four different temperatures (5°C, 15°C, 25°C and 35°C) that were used to assess temperature sensitivity of the activities. The incubation duration at 5°C, 15°C, 25°C and 35°C were 23, 6, 3 and 1.5 h, respectively. After incubation, the plates were centrifuged at 350 g for 3 min, and 250 μL of supernatant from each well was placed into the corresponding well of a black 96-well plate. Fluorescence was measured with 365 and 450 nm in wavelengths for excitation and emission, respectively, using a Tecan Infinite M500 (Tecan Group, Ltd., Männedorf, Switzerland). Linear standard curves, obtained using the MUC or MUB standards, were used to calculate potential enzyme activity for each enzyme and sample. To assess temperature sensitivity of the enzymes, activation energy was calculated using the potential enzyme activities assayed at the four temperatures using the Arrhenius equation (Arrhenius 1889) described in details in Koyama et al. (2013). DNA extraction, PCR and 454 sequencing Subsets of the soil samples were analyzed for the microbial community structure using Roche 454 sequencing. In 2009, soils from nine plots with the control temperature treatment across the three precipitation treatments and three plots with the ambient precipitation and high warming were selected. In 2011, soils from all the plots with the control and high warming treatments across the three precipitation treatments were selected. The selected samples of 2009 and 2011 were processed in the same manner described in details by Evans and Wallenstein (2012) and Koyama et al. (2014), respectively. Briefly, genomic DNA was extracted from each of 0.25 g sub-sample collected from the control and high-temperature treatments using MoBio PowerSoil DNA extraction kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA) and stored at −80°C until further processing. The small subunit of 16S and 18S rRNA genes was amplified for each sample using paired primers of F515/R806 (Bates et al.2010) and SSU817R/SSU1196 (Borneman and Hartin, 2000), respectively. The primers were modified for the 454 sequencing platform (Rousk et al.2010). For the samples collected in 2009, PCR was performed using 50 μL assays; 0.5 μL (10 μM) of each primer, 3 μL (5 ng μL−1) of template DNA, 5 μL of BSA (5 ng μL−1), 18.5 μL of PCR-grade water and 22.5 μL of Platinum PCR SuperMix (Invitrogen, Carlsbad, CA, USA). The PCR thermal profile consisted of an initial denaturation at 94°C for 3 min, followed by 35 cycles of 95°C for 45 s, 50°C for 30 s and 72°C for 90 s with a final extension of 10 min at 72°C. For the samples collected in 2011, PCR was performed using 25 μL assays; 1.25 μL (10 μM) of each primer, 1 μL (1 ng μL−1) of template DNA, 1.25 μL of BSA (10 ng μL−1), 8.5 μL of PCR-grade water, and 12.5 μL of KAPA2G Fast Multiplex Mix (Kapa Biosystems, Woburn, MA, USA). The thermal profile consisted of an initial denaturation at 95°C for 3 min, followed by 30 cycles of 95°C for 10 s, 50°C for 10 s and 72°C for 1 s with a final extension of 10 min at 72°C. Amplicons were evaluated for amplification and lengths by agarose gel electrophoresis, and purified using the UltraClean® PCR Clean-UP Kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA). Appropriate quantities of purified amplicons were pooled and sequenced on a Roche 454 FLX sequencer at Selah Genomics (Greenville, SC, USA). Sequence data processing Sequences were processed via the QIIME 1.9 toolkit (Caporaso et al.2010a). Operational taxonomic units (OTUs) were determined at the ≥97% in similarity of the sequences, and assigned taxonomy via the Ribosomal Database Project (Cole et al.2009) and NCBI BLAST (Johnson et al.2008) for bacteria and fungi, respectively. After chimeric sequences and singletons were removed, total numbers of sequences of the 2009 bacteria, 2011 bacteria and 2011 fungi were 67 706 (ranging from 2513 to 7620 sequences per sample), 63 025 (660 to 8337), and 102 588 (1020 to 4116), respectively. The bacterial sequences collected in 2009 and 2011 were rarefied at 2513 and 660 per sample, respectively, and the fungal sequence at 1020 per sample for downstream analyses. Sequences in each data set were aligned via PyNAST (Caporaso et al.2010b) to build a phylogenetic tree using FastTree (Price, Dehal and Arkin 2009). The downstream analyses included UniFrac (Hamady, Lozupone and Knight 2010) in QIIME, and distant-based redundancy analysis (dbRDA, Legendre and Anderson 1999) using the vegan package in R (R Development Core Team 2015). The sequences were deposited to the MG-RAST server (http://metagenomics.anl.gov/) and are available to the public (accession numbers from 4735126.3 to 4758399.3 under ‘BACE_soil_microbes’). Statistical analyses All the computations were carried out using R (R Development Core Team 2015). Mixed-effect ANOVAs in the lme4 package were employed, with precipitation (i.e. drought, ambient and wet) and temperature (i.e. control, low, medium and high) as fixed effects, and blocks as a random effect. A significance level of P ≤ 0.05 was considered significant. RESULTS Soil properties Both precipitation and temperature treatments significantly altered some soil properties (Fig. 1). Overall, SWC was lowest in the drought and high-temperature treatments (Fig. 1). In 2009, the differences in SWC across the precipitation treatments were widened as the higher heat was applied, supported by a precipitation × temperature interaction effect (P ≤ 0.001, Fig. 1). In the 2011 soils, the drought (P ≤ 0.001, Table 1) and higher temperature treatments (P ≤ 0.001, Table 1) reduced SWC (Fig. 1). The soils under vegetation had lower SWC than bare soils (P ≤ 0.001, Table 1, Fig. 1). Figure 1. View largeDownload slide Properties of soils collected in 2009 and 2011. The soils collected in 2009 were from vegetated areas only. Soils collected in 2011 were from two cover types (vegetated and bare soils). Results of mixed-effect ANOVAs for the 2011 data with the two cover types combined are shown in Table 1. Results of mixed-effect ANOVAs for each cover type are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Figure 1. View largeDownload slide Properties of soils collected in 2009 and 2011. The soils collected in 2009 were from vegetated areas only. Soils collected in 2011 were from two cover types (vegetated and bare soils). Results of mixed-effect ANOVAs for the 2011 data with the two cover types combined are shown in Table 1. Results of mixed-effect ANOVAs for each cover type are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Table 1. Results of mixed-effect ANOVAs for properties of soils collected in June 2011. Soil samples from the two cover types (vegetated and bare soils) were pooled for the analyses. P-values ≤ 0.05 are in bold. Response variables  Independent variables  F values  P values  SWC  Temp  F3,22 = 20.898  P < 0.001    Precip  F2,22 = 24.89  P < 0.001    Cover  F1,24 = 122.294  P < 0.001    Temp × Precip  F6,22 = 1.056  P = 0.418    Temp × Cover  F3,24 = 0.647  P = 0.593    Precip × Cover  F2,24 = 0.993  P = 0.385    Temp × Precip × Cover  F6,24 = 1.188  P = 0.346  SOC  Temp  F3,22 = 0.596  P = 0.625    Precip  F2,22 = 5.825  P = 0.009    Cover  F1,24 = 1.676  P = 0.208    Temp × Precip  F6,22 = 0.572  P = 0.748    Temp × Cover  F3,24 = 0.379  P = 0.769    Precip × Cover  F2,24 = 1.144  P = 0.335    Temp × Precip × Cover  F6,24 = 1.193  P = 0.344  Total N  Temp  F3,22 = 0.321  P = 0.810    Precip  F2,22 = 4.065  P = 0.031    Cover  F1,24 < 0.001  P = 0.987    Temp × Precip  F6,22 = 0.551  P = 0.764    Temp × Cover  F3,24 = 0.384  P = 0.765    Precip × Cover  F2,24 = 0.922  P = 0.412    Temp × Precip × Cover  F6,24 = 1.224  P = 0.329  C:N ratio  Temp  F3,22 = 0.86  P = 0.478    Precip  F2,22 = 7.41  P = 0.004    Cover  F1,24 = 35.35  P < 0.001    Temp × Precip  F6,22 = 0.81  P = 0.570    Temp × Cover  F3,24 = 1.19  P = 0.333    Precip × Cover  F2,24 = 8.13  P = 0.002    Temp × Precip × Cover  F6,24 = 1.69  P = 0.167  SIR  Temp  F3,22 = 4.916  P = 0.009    Precip  F2,22 = 18.952  P < 0.001    Cover  F1,24 = 716.886  P < 0.001    Temp × Precip  F6,22 = 1.109  P = 0.389    Temp × Cover  F3,24 = 1.78  P = 0.178    Precip × Cover  F2,24 = 8.087  P = 0.002    Temp × Precip × Cover  F6,24 = 0.419  P = 0.859  Response variables  Independent variables  F values  P values  SWC  Temp  F3,22 = 20.898  P < 0.001    Precip  F2,22 = 24.89  P < 0.001    Cover  F1,24 = 122.294  P < 0.001    Temp × Precip  F6,22 = 1.056  P = 0.418    Temp × Cover  F3,24 = 0.647  P = 0.593    Precip × Cover  F2,24 = 0.993  P = 0.385    Temp × Precip × Cover  F6,24 = 1.188  P = 0.346  SOC  Temp  F3,22 = 0.596  P = 0.625    Precip  F2,22 = 5.825  P = 0.009    Cover  F1,24 = 1.676  P = 0.208    Temp × Precip  F6,22 = 0.572  P = 0.748    Temp × Cover  F3,24 = 0.379  P = 0.769    Precip × Cover  F2,24 = 1.144  P = 0.335    Temp × Precip × Cover  F6,24 = 1.193  P = 0.344  Total N  Temp  F3,22 = 0.321  P = 0.810    Precip  F2,22 = 4.065  P = 0.031    Cover  F1,24 < 0.001  P = 0.987    Temp × Precip  F6,22 = 0.551  P = 0.764    Temp × Cover  F3,24 = 0.384  P = 0.765    Precip × Cover  F2,24 = 0.922  P = 0.412    Temp × Precip × Cover  F6,24 = 1.224  P = 0.329  C:N ratio  Temp  F3,22 = 0.86  P = 0.478    Precip  F2,22 = 7.41  P = 0.004    Cover  F1,24 = 35.35  P < 0.001    Temp × Precip  F6,22 = 0.81  P = 0.570    Temp × Cover  F3,24 = 1.19  P = 0.333    Precip × Cover  F2,24 = 8.13  P = 0.002    Temp × Precip × Cover  F6,24 = 1.69  P = 0.167  SIR  Temp  F3,22 = 4.916  P = 0.009    Precip  F2,22 = 18.952  P < 0.001    Cover  F1,24 = 716.886  P < 0.001    Temp × Precip  F6,22 = 1.109  P = 0.389    Temp × Cover  F3,24 = 1.78  P = 0.178    Precip × Cover  F2,24 = 8.087  P = 0.002    Temp × Precip × Cover  F6,24 = 0.419  P = 0.859  View Large Contents of SOC were affected by the precipitation treatments. The 2009 vegetated soils had higher SOC content in the drought than the control or wet treatment (P = 0.038, Fig. 1). Overall, a similar trend was found in the soils collected in July 2011 (P = 0.009, Table1, Fig. 1). However, this trend was not significant when only the vegetated soils were examined (P = 0.186). On the other hand, the bare soils in the drought and wet treatments tended to have higher SOC contents than controls (P = 0.011). No treatment significantly altered total N contents in the 2009 vegetated soils (Fig. 1). Overall, total N contents in soils collected in 2011 showed a trend similar to SOC in the same soil samples (Fig. 1) where the precipitation altered total N content (P = 0.009, Table 1). A significant precipitation effect was not detected in the vegetated soils (P = 0.181, Fig. 1), but was found in the bare soils (P = 0.040, Fig. 1). No treatment significantly altered C:N ratios of the 2009 vegetated soils (Fig. 1). The precipitation treatment significantly altered C:N ratio of the 2011 bare soils, but this was not the case for the 2011 vegetated soils (Fig. 1). This was supported by an interaction effect between the precipitation and cover type for the 2011 soils (P = 0.002, Table 1). When the two cover types were compared, C:N ratios were higher in the bare than vegetated soils (P ≤ 0.001, Table 1). In the bare soils, C:N ratios of soils under the drought treatment were highest, followed by those of control and wet precipitation treatments (Fig. 1). No treatment significantly altered SIR in the 2009 vegetated soils (Fig. 1). On the other hand, both temperature and precipitation treatments affected SIR in the 2011 soils (Fig. 1). Soils in the drought treatment showed lowest SIR and the wet treatment showed the highest (P ≤ 0.001, Table 1), and this trend was more pronounced in vegetated than bare soils (Fig. 1). This is supported by a precipitation × cover type effect (P = 0.002, Table 1). The vegetated soils had higher SIR than bare soils (P ≤ 0.001, Table 1). In both types of soils, higher temperature treatments decreased SIR (P = 0.009, Table 1). When SIR data were analyzed for each cover type, both precipitation (P = 0.005 and < 0.001 in vegetated and bare soils, respectively) and temperature treatments (P = 0.023 and 0.021 in vegetated and bare soils, respectively) were significant (Fig. 1). Microbial community composition Bacterial community composition differed among the precipitation treatments in the 2009 vegetated soils, as demonstrated by a UniFrac analysis (P ≤ 0.001 in PC1, Fig. 2). The PC1 values, which explained 49.1% of the variation, co-varied with precipitation quantities, where the drought treatment had the lowest, and the wet treatment the highest values (Fig. 1). When the control and high-temperature treatments were compared for the soils in the ambient precipitation via paired Student t-tests, there were significant differences in both PC1 and PC2 (P = 0.020 and 0.011, respectively). Figure 2. View largeDownload slide Results of UniFrac analyses for bacterial and fungal community composition. Symbols and error bars represent means and standard errors, respectively. Soils collected in 2009 were from vegetated areas subjected to four treatments, and processed for bacterial community composition only. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat) and processed for bacterial and fungal community composition. The UniFrac analyses for the 2011 data were conducted with the two cover types together, but the results of the two cover types were plotted separately to help distinguish the differences. ANOVAs were conducted for PC values of the 2009 soils with the control temperature treatment. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data are shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Figure 2. View largeDownload slide Results of UniFrac analyses for bacterial and fungal community composition. Symbols and error bars represent means and standard errors, respectively. Soils collected in 2009 were from vegetated areas subjected to four treatments, and processed for bacterial community composition only. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat) and processed for bacterial and fungal community composition. The UniFrac analyses for the 2011 data were conducted with the two cover types together, but the results of the two cover types were plotted separately to help distinguish the differences. ANOVAs were conducted for PC values of the 2009 soils with the control temperature treatment. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data are shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. A similar trend was observed for bacterial community composition in the 2011 vegetated soils (Fig. 2). Values of the PC1 co-varied with the precipitation as well as temperature gradients (P < 0.001, and = 0.002, respectively, Fig. 2) when only the PC1 values of the vegetated soils were processed via an ANOVA. The PC2 values were different among the precipitation treatments, where the soils under the ambient treatment had higher scores than the manipulated precipitation treatments (P = 0.026). In contrast to the vegetated soils, there were no precipitation or temperature treatment effects in the bare soils in the PC1 or PC2 values (P ≥ 0.097, Fig. 2). Such contrasting responses of bacterial community composition between the two cover types were supported by a precipitation × cover types effect for the PC1 values (P = 0.002, Table 2). In PC2, most of the values of vegetated and bare soils were positive and negative, respectively (Fig. 2) supported by a significant main cover type effect for the PC2 values (P < 0.001, Table 1). The PC1 and PC2 explained 18.7% and 12.0% of variability in bacterial community structure in the UniFrac analysis (Fig. 2). Table 2. Results of mixed-effect ANOVAs for sample scores of two primary axes resuling from multivariate analyses of bacterial and fungal community composition. Microbial data were obtained from soil samples from the two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat). P-values ≤ 0.05 are shown bold. Response variables  Independent variables  F values  P values  Bacterial weighted UniFrac PC 1  Temp  F1,19 = 5.535  P = 0.030    Precip  F2,19 = 5.555  P = 0.013    Cover  F1,19 = 0.471  P = 0.501    Temp × Precip  F2,19 = 0.146  P = 0.865    Temp × Cover  F1,19 = 1.136  P = 0.300    Precip × Cover  F2,19 = 9.165  P = 0.002    Temp × Precip × Cover  F2,19 = 0.457  P = 0.640  Bacterial weighted UniFrac PC 2  Temp  F1,19 = 6.494  P = 0.020    Precip  F2,19 = 7.693  P = 0.004    Cover  F1,19 = 81.616  P < 0.001    Temp × Precip  F2,19 = 2.54  P = 0.105    Temp × Cover  F1,19 = 0.005  P = 0.944    Precip × Cover  F2,19 = 2.128  P = 0.147    Temp × Precip × Cover  F2,19 = 1.062  P = 0.365  Fungal weighted UniFrac PC 1  Temp  F1,22 = 0.001  P = 0.979    Precip  F2,22 = 0.101  P = 0.905    Cover  F1,22 = 0.261  P = 0.614    Temp × Precip  F2,22 = 0.787  P = 0.468    Temp × Cover  F1,22 = 0.784  P = 0.385    Precip × Cover  F2,22 = 0.485  P = 0.622    Temp × Precip × Cover  F2,22 = 1.143  P = 0.337  Fungal weighted UniFrac PC 2  Temp  F1,22 = 12.535  P = 0.002    Precip  F2,22 = 0.78  P = 0.471    Cover  F1,22 = 0.059  P = 0.811    Temp × Precip  F2,22 = 2.042  P = 0.154    Temp × Cover  F1,22 = 0.141  P = 0.711    Precip × Cover  F2,22 = 5.051  P = 0.016    Temp × Precip × Cover  F2,22 = 0.749  P = 0.485  Bacterial dbRDA Axis 1  Temp  F1,19 = 10.074  P = 0.005    Precip  F2,19 = 7.016  P = 0.005    Cover  F1,19 = 6.128  P = 0.023    Temp × Precip  F2,19 = 0.069  P = 0.933    Temp × Cover  F1,19 = 0.547  P = 0.468    Precip × Cover  F2,19 = 10.373  P = 0.001    Temp × Precip × Cover  F2,19 = 0.28  P = 0.759  Bacterial dbRDA Axis 2  Temp  F1,19 = 0.032  P = 0.860    Precip  F2,19 = 1.377  P = 0.277    Cover  F1,19 = 19.613  P < 0.001    Temp × Precip  F2,19 = 3.217  P = 0.063    Temp × Cover  F1,19 = 1.135  P = 0.300    Precip × Cover  F2,19 = 9.765  P = 0.001    Temp × Precip × Cover  F2,91 = 3.932  P = 0.037  Response variables  Independent variables  F values  P values  Bacterial weighted UniFrac PC 1  Temp  F1,19 = 5.535  P = 0.030    Precip  F2,19 = 5.555  P = 0.013    Cover  F1,19 = 0.471  P = 0.501    Temp × Precip  F2,19 = 0.146  P = 0.865    Temp × Cover  F1,19 = 1.136  P = 0.300    Precip × Cover  F2,19 = 9.165  P = 0.002    Temp × Precip × Cover  F2,19 = 0.457  P = 0.640  Bacterial weighted UniFrac PC 2  Temp  F1,19 = 6.494  P = 0.020    Precip  F2,19 = 7.693  P = 0.004    Cover  F1,19 = 81.616  P < 0.001    Temp × Precip  F2,19 = 2.54  P = 0.105    Temp × Cover  F1,19 = 0.005  P = 0.944    Precip × Cover  F2,19 = 2.128  P = 0.147    Temp × Precip × Cover  F2,19 = 1.062  P = 0.365  Fungal weighted UniFrac PC 1  Temp  F1,22 = 0.001  P = 0.979    Precip  F2,22 = 0.101  P = 0.905    Cover  F1,22 = 0.261  P = 0.614    Temp × Precip  F2,22 = 0.787  P = 0.468    Temp × Cover  F1,22 = 0.784  P = 0.385    Precip × Cover  F2,22 = 0.485  P = 0.622    Temp × Precip × Cover  F2,22 = 1.143  P = 0.337  Fungal weighted UniFrac PC 2  Temp  F1,22 = 12.535  P = 0.002    Precip  F2,22 = 0.78  P = 0.471    Cover  F1,22 = 0.059  P = 0.811    Temp × Precip  F2,22 = 2.042  P = 0.154    Temp × Cover  F1,22 = 0.141  P = 0.711    Precip × Cover  F2,22 = 5.051  P = 0.016    Temp × Precip × Cover  F2,22 = 0.749  P = 0.485  Bacterial dbRDA Axis 1  Temp  F1,19 = 10.074  P = 0.005    Precip  F2,19 = 7.016  P = 0.005    Cover  F1,19 = 6.128  P = 0.023    Temp × Precip  F2,19 = 0.069  P = 0.933    Temp × Cover  F1,19 = 0.547  P = 0.468    Precip × Cover  F2,19 = 10.373  P = 0.001    Temp × Precip × Cover  F2,19 = 0.28  P = 0.759  Bacterial dbRDA Axis 2  Temp  F1,19 = 0.032  P = 0.860    Precip  F2,19 = 1.377  P = 0.277    Cover  F1,19 = 19.613  P < 0.001    Temp × Precip  F2,19 = 3.217  P = 0.063    Temp × Cover  F1,19 = 1.135  P = 0.300    Precip × Cover  F2,19 = 9.765  P = 0.001    Temp × Precip × Cover  F2,91 = 3.932  P = 0.037  View Large When the bacterial community composition was further examined at a higher taxonomic level using dbRDA, similar trends were found (Fig. 3). The axis 1 values of the 2009 soils collected under vegetation, which explained 53.9% of the variability, co-varied with water availability (P < 0.001, Fig. 3). The primary driver of the co-variation was due to the dominance of Actinobacteria in the drought treatment (Fig. 3; Fig. S1, Supporting Information). Along the axis 2, the drought and wet treatments had lower values than the ambient treatment (P = 0.003, Fig. 3). Figure 3. View largeDownload slide Results of distance-based redundancy analyses (dbRDA) for the composition of bacterial taxa in soils collected in 2009 and 2011. Soils collected in 2009 were from vegetated areas subjected to four treatments. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat). The dbRDA for the 2011 data were conducted with the two cover types combined, but the results of the two cover types were plotted separately to help distinguish the differences. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data were shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Proteo; Proteobacteria, Acido: Acidobacteria. Figure 3. View largeDownload slide Results of distance-based redundancy analyses (dbRDA) for the composition of bacterial taxa in soils collected in 2009 and 2011. Soils collected in 2009 were from vegetated areas subjected to four treatments. Soils collected in 2011 were from two cover types (vegetated and bare soils) subjected to two temperature treatments (control and high heat). The dbRDA for the 2011 data were conducted with the two cover types combined, but the results of the two cover types were plotted separately to help distinguish the differences. Results of mixed-effect ANOVAs with the two cover types combined for the 2011 data were shown in Table 2. Results of mixed-effect ANOVAs for each cover type and axis are shown in corresponding panels (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Proteo; Proteobacteria, Acido: Acidobacteria. Bacterial community composition in the 2011 vegetated soils via dbRDA showed a similar trend to the UniFrac analysis (Fig. 3). The axis 1 values co-varied with the precipitation as well as temperature gradients supported by significant precipitation and temperature effects (P < 0.001, Fig. 3) when only the axis 1 values of the vegetated soils were analyzed via an ANOVA. This co-variation was primarily driven by the dominance of Actinobacteria in the drought and high-temperature treatments (Figs 3 and 4). In contrast to the vegetated soils, there were no significant differences among the precipitation or temperature treatments in the bare soils in the axis 1 values (P ≥ 0.189, Fig. 3). Such contrasting responses of bacterial community composition between the two cover types were supported by a precipitation × cover type effect as well as a main cover type effect for the axis 1 values (P = 0.001 and 0.023, respectively, Table 2). The axis 2 values of the vegetated soils were different among the precipitation treatments, where the soils under the ambient treatment had higher scores than the manipulated precipitation treatments (P = 0.026, Fig. 3). The axis 2 values of the bare soils indicated the two treatments significantly altered relative abundances of bacterial taxa, supported by temperature × precipitation and main precipitation effects (P = 0.05 and 0.036, respectively, Fig. 3). Overall, the vegetated soils had higher values than the bare soils along axis 2 (Fig. 3). This difference was supported by precipitation × cover type and main cover type effects (P = 0.001 and < 0.001, respectively, Table 1). The axes 1 and 2, respectively, explained 32.9% and 8.3% of variability in the bacterial community structure in the dbRDA analysis for the 2011 soils (Fig. 3). Figure 4. View largeDownload slide Mean relative abundances of major bacterial taxa in the vegetated and bare soils collected in 2011. Drought, ambient and wet represent three levels of the precipitation treatments. C and H represent the control and high-temperature treatments, respectively. Figure 4. View largeDownload slide Mean relative abundances of major bacterial taxa in the vegetated and bare soils collected in 2011. Drought, ambient and wet represent three levels of the precipitation treatments. C and H represent the control and high-temperature treatments, respectively. The fungal community structure was less affected by the treatments relative to that of bacteria (Fig. 2). When both cover types were analyzed together, there was no significant treatment effect on PC1, which explained 48.9% of variability (Table 2, Fig. 2). In PC2, which explained 11.9% of variability, the temperature treatment was significant (P = 0.002, Table 2) where the values of control tended to be higher than those of the high-temperature treatment when paired by the precipitation treatments for each cover type (Fig. 2). There was a precipitation × cover type effect for PC2 (P = 0.016, Table 2). This was likely caused by the significant precipitation effect in the vegetated soils but a lack of this trend in the bare soils when the cover types were analyzed separately (Fig. 2). Due to the lack of clear responses to the treatments and low taxonomic resolution (Figs S2 and S3, Supporting Information), fungal relative abundances at higher taxonomic ranks were not analyzed further. Enzyme assay Overall, four enzymes, BG, CB, LAP and PHOS, had higher potential activities when more precipitation was applied and had lower activities when increasing heat was applied in the 2011 vegetated soils (Fig. 5). Potential enzyme activities of XYL were significantly decreased with increasing heat application (Fig. 5). No significant alteration by the treatments was observed for AG or NAG (Fig. 5). Overall, potential activities of the seven enzymes showed similar results when assessed at different temperatures (Fig. 5; Fig. S4, Supporting Information). Figure 5. View largeDownload slide Potential enzyme activities (μmol g−1 h−1) of seven different enzymes assayed at 15°C, activation energy (kJ mol−1) of the seven enzymes, and stoichiometry (unit-less) of potential enzyme activities assayed at 15°C for the soils collected from vegetated areas in 2011. The activation energy was calculated using potential enzyme activities assayed at 5°C, 15°C, 25°C and 35°C (Fig. S5, Supporting Information). Results of mixed-effect ANOVAs are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Figure 5. View largeDownload slide Potential enzyme activities (μmol g−1 h−1) of seven different enzymes assayed at 15°C, activation energy (kJ mol−1) of the seven enzymes, and stoichiometry (unit-less) of potential enzyme activities assayed at 15°C for the soils collected from vegetated areas in 2011. The activation energy was calculated using potential enzyme activities assayed at 5°C, 15°C, 25°C and 35°C (Fig. S5, Supporting Information). Results of mixed-effect ANOVAs are shown in corresponding panels. T and P represent temperature and precipitation treatments, respectively. *P ≤ 0.05, **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Increasing precipitation reduced activation energy of NAG in all temperature treatments and that of CB only in the control temperature (Fig. 5) supported by a temperature × precipitation effect (P = 0.022). Activation energy of PHOS was higher for the drought and wet treatments than the ambient, as supported by a precipitation effect (P = 0.012, Fig. 5). Increasing precipitation had higher BG:N and C:N ratios, and lower N:P ratios (P = 0.007, 0.018 and 0.034, respectively, Fig. 5). Ratios of BG:P and C:P were lower in the drought and wet treatments than the control (P = 0.007 for both ratios, Fig. 5). A redundancy analysis was conducted to analyze the relative abundances of the major bacterial taxa (Figs 3 and 4) and the activation energy and stoichiometry significantly altered by the precipitation and/or temperature treatments in the 2011 vegetated soils (Fig. 6). Primary sample scores were similar to those in UniFrac (Fig. 2) and dbRDA (Fig. 3), where the scores in the axis 1 co-varied with precipitation quantities (P ≤ 0.001), and increasing temperature decreased the axis 1 scores (P = 0.005). These treatment effects were driven by an increasing relative abundance of Actinobacteria with decreasing precipitation quantities and with warming (Fig. 6). The higher relative abundance of Actinobacteria was associated with increasing NAG activation energy, and N:P and C:P enzyme ratios, and with decreasing CB activation energy and C:N enzyme ratio (Fig. 6). No significant treatment effect was found for the axis 2 scores (Fig. 6). Axes 1 and 2 explained 35.1% of the variability of the data (Fig. 6). Figure 6. View largeDownload slide Results of redundancy analysis using composition of bacterial taxa in the soils collected from vegetated areas in 2011, and the enzyme activation energy and stoichiometry values that were found significantly affected by the experimental treatments in the mixed-effect ANOVAs (Fig. 4). The result of mixed-effect ANOVA for each axis of sample scores is shown in the corresponding panel (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Bactero: Bacteroidetes, Gemmati: Gemmatimonadetes, Proteo; Proteobacteria Figure 6. View largeDownload slide Results of redundancy analysis using composition of bacterial taxa in the soils collected from vegetated areas in 2011, and the enzyme activation energy and stoichiometry values that were found significantly affected by the experimental treatments in the mixed-effect ANOVAs (Fig. 4). The result of mixed-effect ANOVA for each axis of sample scores is shown in the corresponding panel (lower right and upper left for axis 1 and 2, respectively). T and P represent temperature and precipitation treatments, respectively. **P ≤ 0.01, P ≤ 0.001. Where there are no letters, no significant temperature or precipitation effects were found. Bactero: Bacteroidetes, Gemmati: Gemmatimonadetes, Proteo; Proteobacteria DISCUSSION Manipulations of precipitation and temperature significantly altered the soil bacterial community structure in the vegetated plots, and some biogeochemical properties co-varied with the bacterial structural change. However, the climate manipulations had little impact on the bacterial community structure in bare soils despite significant cumulative treatment effects on soil properties including the C:N ratio. The treatment effects on the fungal community structure were not apparent in either vegetated or bare soils. Our results emphasize that the primary responses of soil bacterial communities to changes in precipitation and temperature were mediated through responses of the vascular plant community, as opposed to being direct responses to the changes in abiotic conditions. Higher relative abundances of Actinobacteria in vegetated soils in less-watered treatments were the dominant change in the bacterial community structure in both 2009 and 2011. The Actinobacteria phylum has many members known to be resistant to drought and able to grow under dry conditions (Goodfellow and Williams 1983; Zvyagintsev et al.2007); some studies have reported higher relative abundances of Actinobacteria in dry soil conditions (Cruz-Martínez et al.2012; Barnard, Osborne and Firestone 2013). Actinobacteria are Gram-positive bacteria, which possess stronger cell walls than Gram-negative bacteria, and are thus considered resistant to desiccation (Schimel, Balser and Wallenstein 2007). Some Actinobacteria members can produce stress-resistant spores (Ait Barka et al.2016), which can help them persist in drought (Santos-Medellin et al.2017; Taketani et al.2017, but see Naylor et al.2017). These characteristics might contribute to the high relative abundances of Actinobacteria in less-watered soils in this study. The co-variation between relative abundances of Actinobacteria and the precipitation quantities in this study appears to have been mediated by the responses of the vascular plant community to the precipitation treatments, given that no such trend was observed in the bare soils in 2011 (Fig. 3). This contrasting precipitation effect between the vegetated and bare soils is consistent with two recent reports by Naylor et al. (2017) and Santos-Medellin et al. (2017); both studies demonstrated that drought significantly increased relative abundances of Actinobacteria in rhizosphere soils, but such trends were limited in bulk soils. One plausible mechanism behind the co-variation was conserved common responses of vascular plant species to drought (Naylor et al.2017). This is supported by two studies showing that increased relative abundances of Actinobacteria in the rhizosphere as well as the root endosphere induced by drought were observed consistently in four rice cultivars (Santos-Medellin et al.2017), and 18 plant species belonging to the Poaceae family in addition to tomato, which was used as an outgoup (Naylor et al.2017). Thus, it is possible that many plant species in the less-watered plots in this study increased production of biochemical compounds in response to water stress, which in turn increased relative Actinobacteria abundance. Another potential mechanism contributing to the consistent co-variation was altered plant community structure. At the onset of the experiment in 2008, the old field experimental site was dominated by C3 perennials (Hoeppner and Dukes 2012). By 2010, non-perennials and C4 plants became more common in the ambient and wet than the drought treatment (Hoeppner and Dukes 2012). Plant species differentially shape rhizosphere properties via root exudates (Hertenberger, Zampach and Bachmann 2002), litter quality and quantity (de Deyn, Cornelissen and Bardgett 2008; Meier and Bowman 2008), nutrient uptakes (Bell et al.2015), mycorrhizal association (Rillig and Mummey 2006), pH (Herr et al.2007) and soil aggregates (Haynes and Beare 1997). The differences in plant community structure among the precipitation and temperature treatments could have influenced rhizosphere properties, which in turn resulted in differences in the microbial community structure (el Zahar Haichar et al.2008; Philippot et al.2013; Tkacz et al.2015). Nitrogen dynamics under vegetation could also be a driver for the observed differences in the bacterial community structure. Inorganic N amendment to mineral soils diverse in origin consistently increased relative abundances of Actinobacteria (Ramirez, Craine and Fierer 2012). This result was supported by a greenhouse experiment conducted by Bell et al. (2015) who demonstrated that relative abundances of Actinobacteria consistently declined during the growing season as available N in soils decreased due to N assimilation by four different grass species. In our study, indeed, inorganic N concentrations in vegetated soils of 2011 were higher in the drought than ambient or wet precipitation treatments (Fig. S5, Supporting Information), and relative abundances of Actinobacteria were significantly correlated with inorganic N concentration (Fig. S6, Supporting Information). Relative quantities of available C altered by the response of plant community to the precipitation treatments might not be a contributor to the observed microbial community shifts, for two reasons. First, increased available C should, in theory, favor copiotrophs over oligotrophs, but the phylum Actinobacteria may not be classified as either of the categories (Fierer, Bradford and Jackson 2007). Second, net primary productivity, which should affect available C in the vegetated soils, was not significantly altered by the precipitation treatments (Hoeppner and Dukes 2012). Differences in available C should be found, instead, between the vegetated and bare soils; we hypothesized that copiotrophs should be favored in the vegetated soils that were supplied with C via rhizodeposition from the vegetation, and oligotrophs were favored in bare soils that had less available C without organic matter supply from vascular plants. Our results supported this hypothesis. The vegetated and bare soils collected in 2011 had higher relative abundances of β-Proteobacteria and Acidobacteria, respectively (Fig. 3), which are considered as copiotrophic and oligotrophic taxa (Fierer, Bradford and Jackson 2007). The constituents of rhizodeposition, such as root exudates, are C-rich low molecular-weight compounds released by living roots (Hütsch, Augustin and Merbach 2002), and are immediately assimilated as a C source by rhizosphere microbes upon production (Denef et al.2009), and thus should support copiotrophs over oligotrophs. We do not know the mechanisms behind the higher relative abundances of α-Proteobacteria in the bare than the vegetated soils (Fig. 2). This observation contrasts with the higher α-Proteobacteria in vegetated than bare soils reported by Thomson et al. (2010) and Yergeau et al. (2012). It is possible that relative abundance of α-Proteobacteria was passively increased in the bare soils due to the increased relative abundances of copiotrophs such as β-Proteobacteria and Bacteroidetes (Fierer, Bradford and Jackson 2007) under vegetation. Mycorrhizal fungi associated with the plant community could interact with the rhizosphere bacterial community (Bonfante and Anca 2009). Nuccio et al. (2013) reported that the presence of arbuscular mycorrhizal fungi (phylum Glomeromycota) reduced relative abundances of Actinobacteria. Our data of relative abundance of Glomeromycota show a trend consistent with potential influence of Glomeromycota on the Actinobacteria population in the high-temperature treatment, but not in the control (Fig. S7, Supporting Information). It is possible that absolute abundances of Glomeromycota were positively correlated with quantities of precipitation, which in turn affected the relative abundance of Actinobacteria. The significant temperature effect on bacterial community structure in the 2011 soils was also mediated by the presence of plants (Figs 2 and 3). The lack of temperature effects on bacterial community structure in the bare soils is consistent with results of Oliverio, Bradford and Fierer (2017) who investigated bacterial response to increased temperature in bare soils. In the study, a group of bacterial OTUs consistently responded to increased temperature in a lab incubation in bare soils collected from diverse ecosystems. However, the OTUs did not appear to be phylogenetically predictable, consistent with our results of no significant response to increased temperature in the bare soils using UniFrac (Fig. 2) or dbRDA (Fig. 3). Thus, it appears that the soil microbial community structure is not directly sensitive to ecosystem warming. Five of the seven potential enzyme activities assessed were significantly affected by the precipitation and/or temperature treatments, where BG, CB, LAP and PHOS showed a similar trend; the potential activities were lower in plots with lower precipitation and higher temperature (Fig. 5). One of the primary factors affecting potential enzyme activities is population sizes of soil microbes that produce extracellular enzymes (Nannipieri, Kandeler and Ruggiero 2002); thus microbial biomass can be positively correlated with potential enzyme activities (e.g. Waldrop, McColl and Powers 2003; Tan, Chang and Kabzems 2008; Keeler, Hobbie and Kellog 2009; Brockett, Prescott and Grayston 2012, but see Bell, Klironomos and Henry 2010). The four potential enzyme activities were, indeed, positively correlated with substrate induced respiration (Fig. S8, Supporting Information), an index of microbial biomass (Anderson and Domsch 1978). On the other hand, properties of the potential enzyme activities, including temperature sensitivity and stoichiometry, should reflect microbial community composition, which controls enzyme qualities (Alster et al. 2016a, b). We found that four enzyme properties co-varied with the primary axis reflecting microbial community structure (Fig. 6). Though the co-variations are only correlations, they indicate that differences in bacterial community structure could affect the extracellular enzyme properties in the vegetated soils. This result makes this one of the few studies to link microbial community structure and soil extracellular enzyme activities (Bier et al.2015). We do not know the mechanisms behind our observation that the fungal community structure did not respond to the precipitation or temperature treatments with the same magnitude as the bacterial community (Fig. 2). Reponses of bacterial and fungal community structure to environmental manipulations, such as precipitation and temperature, can depend on the type of ecosystem and manipulation, as various results have been reported. In responses to precipitation manipulations, for instance, some studies showed trends similar to this study with a significant responses of community structure for bacteria, but to lesser degrees for fungi (Barnard, Osborne and Firestone 2013; Amend et al.2016; Zhang et al.2016; Naylor et al.2017). However, some studies reported different trends, including an opposite trend with little effect on bacteria but stronger effects on fungi (Castro et al.2010), bacteria and fungi both being significantly affected (McHugh and Schwartz 2015; Waring and Hawkes 2015; Santos-Medellin et al.2017), and neither of them being altered (McHugh and Schwartz 2015). In addition to precipitation quantities, timing of precipitation manipulations can be an important factor; in a Mediterranean-type grassland, increased precipitation during rainy seasons significantly altered a plant community compared to controls (Suttle, Thomsen and Power 2007), which was accompanied with significant changes in the fungal community structure in soils (Hawkes et al.2011) but not in the bacterial community structure (Cruz-Martí et al.2009). In contrast, in a different Mediterranean-type grassland, the bacterial community structure was significantly changed, but fungal community was not in response to water addition during the dry season (Barnard, Osborne and Firestone 2013). These contrasting responses of bacteria vs. fungi to precipitation manipulations make it a challenge to generalize responses of the two groups to environmental changes (Griffiths and Philippot 2013). Another potential explanation for the lack of fungal community response to the treatments in this study could be the fungal 18S primer set employed in this study, as primer selectivity can cause quantitative biases in amplicons (Peršoh 2015). However, Barnard et al. (2013) and Naylor et al. (2017) used different sets of fungal-specific primers, targeting the 28S and ITS2 regions, respectively, and found little drought effect on a rhizosphere fungal community, whereas responses of the co-occurring rhizosphere bacterial community were similar to those of this study. In conclusion, our results demonstrated that soil microbial communities, which play a major role in C and nutrient biogeochemistry, respond to environmental changes primarily through shifts in plant-soil-microbial interactions, as opposed to responding directly to the changes in abiotic conditions. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. Acknowledgements We thank Drs Guy Beresford and Ann M. Hess for consultation in molecular work and statistics, respectively. We thank Carol Goranson for help at the BACE site. FUNDING This work was supported by grants from the Office of Biological and Environmental Research (BER) of the U.S. Department of Energy's Office of Science to M.D.W. and J.S.D. through the Northeastern Regional Center of the National Institute for Climate Change Research, and from the National Science Foundation (Division of Environmental Biology 0546670 to J.S.D.). Additional support for J.S.D.'s participation in this project was provided by Hatch project 1000026 of the United States Department of Agriculture's National Institute of Food and Agriculture. Conflicts of interest. None declared. REFERENCES Ait Barka E, Vatsa P, Sanchez L et al.   Taxonomy, physiology, and natural products of Actinobacteria. Microbiol Mol Biol Rev  2016; 80: 1– 43. Google Scholar CrossRef Search ADS PubMed  Alster CJ, Baas P, Wallenstein MD et al.   Temperature sensitivity as a microbial trait using parameters from macromolecular rate theory. FrontMicrobiol  2016a; 7: 1– 10. Alster CJ, Koyama A, Johnson NG et al.   Temperature sensitivity of soil microbial communities: an application of macromolecular rate theory to microbial respiration. J Geophys Res:Biogeosci  2016b; 1420– 1433. Amend AS, Martiny AC, Allison SD et al.   Microbial response to simulated global change is phylogenetically conserved and linked with functional potential. ISME J  2016; 10: 109– 18. Google Scholar CrossRef Search ADS PubMed  Anderson JPE, Domsch KH. A physiological method for the quantitative measurement of microbial biomass in soils. Soil Biol Biochem  1978; 10: 215– 21. Google Scholar CrossRef Search ADS   Arrhenius S. Über die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Säuren. Z Phys Chem  1889; 4: 226– 48. Averill C, Turner BL, Finzi AC. Mycorrhiza-mediated competition between plants and decomposers drives soil carbon storage. Nature  2014; 505: 543– 5. Google Scholar CrossRef Search ADS PubMed  Barnard RL, Osborne CA, Firestone MK. Responses of soil bacterial and fungal communities to extreme desiccation and rewetting. ISME J  2013; 7: 2229– 41. Google Scholar CrossRef Search ADS PubMed  Bates ST, Berg-Lyons D, Caporaso JG et al.   Examining the global distribution of dominant archaeal populations in soil. ISME J  2010; 5: 908– 17. Google Scholar CrossRef Search ADS PubMed  Bell CW, Asao S, Calderon F et al.   Plant nitrogen uptake drives rhizosphere bacterial community assembly during plant growth. Soil Biol Biochem  2015; 85: 170– 82. Google Scholar CrossRef Search ADS   Bell TH, Klironomos JN, Henry HA. Seasonal responses of extracellular enzyme activity and microbial biomass to warming and nitrogen addition. Soil Sci Soc Am J  2010; 74: 820– 8. Google Scholar CrossRef Search ADS   Bier RL, Bernhardt ES, Boot CM et al.   Linking microbial community structure and microbial processes: an empirical and conceptual overview. FEMS Microbiol Ecol  2015; 91: fiv113. Google Scholar CrossRef Search ADS PubMed  Blagodatskaya E, Blagodatsky S, Khomyakov N et al.   Temperature sensitivity and enzymatic mechanisms of soil organic matter decomposition along an altitudinal gradient on Mount Kilimanjaro. Sci Rep  2016; 6: 1– 11. Google Scholar CrossRef Search ADS PubMed  Bonfante P, Anca I-A. Plants, mycorrhizal fungi, and bacteria: a network of interactions. Annu Rev Microbiol  2009; 63: 363– 83. Google Scholar CrossRef Search ADS PubMed  Borneman J, Hartin RJ. PCR primers that amplify fungal rRNA genes from environmental samples. Appl Environ Microbiol  2000; 66: 4356– 4360. Google Scholar CrossRef Search ADS PubMed  Brockett BFT, Prescott CE, Grayston SJ. Soil moisture is the major factor influencing microbial community structure and enzyme activities across seven biogeoclimatic zones in western Canada. Soil Biol Biochem  2012; 44: 9– 20. Google Scholar CrossRef Search ADS   Cai W, Borlace S, Lengaigne M et al.   Increasing frequency of extreme El Nino events due to greenhouse warming. Nature Clim Change  2014; 4: 111– 6. Google Scholar CrossRef Search ADS   Campbell CD, Chapman SJ, Cameron CM et al.   A rapid microtiter plate method to measure carbon dioxide evolved from carbon substrate amendments so as to determine the physiological profiles of soil microbial communities by using whole soil. Appl Environ Microbiol  2003; 69: 3593– 9. Google Scholar CrossRef Search ADS PubMed  Caporaso JG, Bittinger K, Bushman FD et al.   PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics  2010b; 26: 266– 7. Google Scholar CrossRef Search ADS   Caporaso JG, Kuczynski J, Stombaugh J et al.   QIIME allows analysis of high-throughput community sequencing data. Nat Methods  2010a; 7: 335– 6. Google Scholar CrossRef Search ADS   Castro HF, Classen AT, Austin EE et al.   Soil microbial community responses to multiple experimental climate change drivers. Appl Environ Microb  2010; 76: 999– 1007. Google Scholar CrossRef Search ADS   Cole JR, Wang Q, Cardenas E et al.   The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res  2009; 37: D141– 5. Google Scholar CrossRef Search ADS PubMed  Chen I-C, Hill JK, Ohlemüller R et al.   Rapid range shifts of species associated with high levels of climate warming. Science  2011; 333: 1024– 6. Google Scholar CrossRef Search ADS PubMed  Cotrufo MF, Soong JL, Horton AJ et al.   Formation of soil organic matter via biochemical and physical pathways of litter mass loss. Nat Geosci  2015; 8: 776. Google Scholar CrossRef Search ADS   Cowles JM, Wragg PD, Wright AJ et al.   Shifting grassland plant community structure drives positive interactive effects of warming and diversity on aboveground net primary productivity. Glob Change Biol  2016; 22: 741– 9. Google Scholar CrossRef Search ADS   Cruz-Martínez K, Rosling A, Zhang Y et al.   Effect of rainfall-induced soil geochemistry dynamics on grassland soil microbial communities. Appl Environ Microb  2012; 78: 7587– 95. Google Scholar CrossRef Search ADS   Cruz-Martinez K, Suttle KB, Brodie EL et al.   Despite strong seasonal responses, soil microbial consortia are more resilient to long-term changes in rainfall than overlying grassland. ISME J  2009; 3: 738– 44. Google Scholar CrossRef Search ADS PubMed  de Deyn GB, Cornelissen JHC, Bardgett RD. Plant functional traits and soil carbon sequestration in contrasting biomes. Ecol Lett  2008; 11: 516– 31. Google Scholar CrossRef Search ADS PubMed  Denef K, Roobroeck D, Manimel Wadu MCW et al.   Microbial community composition and rhizodeposit-carbon assimilation in differently managed temperate grassland soils. Soil Biol Biochem  2009; 41: 144– 53. Google Scholar CrossRef Search ADS   el Zahar Haichar F, Marol C, Berge O et al.   Plant host habitat and root exudates shape soil bacterial community structure. ISME J  2008; 2: 1221– 30. Google Scholar CrossRef Search ADS PubMed  Evans SE, Wallenstein MD. Soil microbial community response to drying and rewetting stress: does historical precipitation regime matter? Biogeochemistry  2012; 109: 101– 16. Google Scholar CrossRef Search ADS   Falkowski PG, Fenchel T, Delong EF. The microbial engines that drive Earth's biogeochemical cycles. Science  2008; 320: 1034– 9. Google Scholar CrossRef Search ADS PubMed  Feng X, Simpson MJ. Temperature and substrate controls on microbial phospholipid fatty acid composition during incubation of grassland soils contrasting in organic matter quality. Soil Biol Biochem  2009; 41: 804– 12. Google Scholar CrossRef Search ADS   Fierer N, Bradford MA, Jackson RB. Toward an ecological classification of soil bacteria. Ecology  2007; 88: 1354– 64. Google Scholar CrossRef Search ADS PubMed  Fierer N, Jackson RB. The diversity and biogeography of soil bacterial communities. P Natl Acad Sci USA  2006; 103: 626– 31. Google Scholar CrossRef Search ADS   Gibbons SM, Lekberg Y, Mummey DL et al.   Invasive plants rapidly reshape soil properties in a grassland ecosystem. mSystems  2017; 2: 1– 13. Google Scholar CrossRef Search ADS   Goodfellow M, Williams ST. Ecology of Actinomycetes. Annu Rev Microbiol  1983; 37: 189– 216. Google Scholar CrossRef Search ADS PubMed  Griffiths BS, Philippot L. Insights into the resistance and resilience of the soil microbial community. FEMS Microbiol Rev  2013; 37: 112– 29. Google Scholar CrossRef Search ADS PubMed  Hamady M, Lozupone C, Knight R. Fast UniFrac: facilitating high-throughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J  2010; 4: 17– 27. Google Scholar CrossRef Search ADS PubMed  Haynes R, Beare M. Influence of six crop species on aggregate stability and some labile organic matter fractions. Soil Biol Biochem  1997; 29: 1647– 53. Google Scholar CrossRef Search ADS   Hawkes CV, Kivlin SN, Rocca JD et al.   Fungal community responses to precipitation. Glob Change Biol  2011; 17: 1637– 45. Google Scholar CrossRef Search ADS   Herr C, Chapuis, LL, Dassonville N et al.   Seasonal effect of the exotic invasive plant Solidago gigantea on soil pH and P fractions. J Plant Nutr Soil Sci  2007; 170: 729– 38. Google Scholar CrossRef Search ADS   Hertenberger G, Zampach P, Bachmann G. Plant species affect the concentration of free sugars and free amino acids in different types of soil. J Plant Nutr Soil Sci  2002; 165: 557– 65. Google Scholar CrossRef Search ADS   Hoeppner SS, Dukes JS. Interactive responses of old-field plant growth and composition to warming and precipitation. Global Change Biol  2012; 18: 1754– 68. Google Scholar CrossRef Search ADS   Hütsch BW, Augustin J, Merbach W. Plant rhizodeposition —an important source for carbon turnover in soils. J Plant Nutr Soil Sci  2002; 165: 397– 407. Google Scholar CrossRef Search ADS   Johnson M, Zaretskaya I, Raytselis Y et al.   NCBI BLAST: a better web interface. Nucleic Acids Research  2008; 36: W5– W9. Google Scholar CrossRef Search ADS PubMed  Judd CR, Koyama A, Simmons MP et al.   Co-variation in methanotroph community composition and activity in three temperate grassland soils. Soil Biol Biochem  2016; 95: 78– 86. Google Scholar CrossRef Search ADS   Kardol P, Campany CE, Souza L et al.   Climate change effects on plant biomass alter dominance patterns and community evenness in an experimental old‐field ecosystem. Global Change Biol  2010; 16: 2676– 87. Google Scholar CrossRef Search ADS   Keeler B, Hobbie S, Kellogg L. Effects of long-term nitrogen addition on microbial enzyme activity in eight forested and grassland sites: implications for litter and soil organic matter decomposition. Ecosystems  2009; 12: 1– 15. Google Scholar CrossRef Search ADS   Kelly AE, Goulden ML. Rapid shifts in plant distribution with recent climate change. P Natl Acad Sci  2008; 105: 11823– 6. Google Scholar CrossRef Search ADS   Koyama A, Wallenstein MD, Simpson RT et al.   Carbon-degrading enzyme activities stimulated by increased nutrient availability in arctic tundra soils. PLoS One  2013; 8: e77212. Google Scholar CrossRef Search ADS PubMed  Koyama A, Wallenstein MD, Simpson RT et al.   Soil bacterial community composition altered by increased nutrient availability in Arctic tundra soils. Front Microbiol  2014; 5: 1– 16. Google Scholar CrossRef Search ADS PubMed  Lauber CL, Hamady M, Knight R et al.   Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl Environ Microb  2009; 75: 5111– 20. Google Scholar CrossRef Search ADS   Legay N, Baxendale C, Grigulis K et al.   Contribution of above-and below-ground plant traits to the structure and function of grassland soil microbial communities. Ann Bot  2014; 114: 1011– 21. Google Scholar CrossRef Search ADS PubMed  Legendre P, Anderson MJ. Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecol Monogr  1999; 69: 1– 24. Google Scholar CrossRef Search ADS   Levine UY, Teal TK, Robertson GP et al.   Agriculture's impact on microbial diversity and associated fluxes of carbon dioxide and methane. ISME J  2011; 5: 1683– 91. Google Scholar CrossRef Search ADS PubMed  Louca S, Jacques SMS, Pires APF et al.   High taxonomic variability despite stable functional structure across microbial communities. Nat Ecol Evol  2017; 1: 0015. Google Scholar CrossRef Search ADS   Makino W, Cotner J, Sterner R et al.   Are bacteria more like plants or animals? Growth rate and resource dependence of bacterial C: N: P stoichiometry. Funct Ecol  2003; 17: 121– 30. Google Scholar CrossRef Search ADS   McHugh TA, Schwartz E. Changes in plant community composition and reduced precipitation have limited effects on the structure of soil bacterial and fungal communities present in a semiarid grassland. Plant Soil  2015; 388: 175– 86. Google Scholar CrossRef Search ADS   Meier CL, Bowman WD. Links between plant litter chemistry, species diversity, and below-ground ecosystem function. P Natl Acad Sci  2008; 105: 19780– 5. Google Scholar CrossRef Search ADS   Min S-K, Zhang X, Zwiers FW et al.   Human contribution to more-intense precipitation extremes. Nature  2011; 470: 378– 81. Google Scholar CrossRef Search ADS PubMed  Moore JC, Berlow EL, Coleman DC et al.   Detritus, trophic dynamics and biodiversity. Ecol Lett  2004; 7: 584– 600. Google Scholar CrossRef Search ADS   Mueller K, Blumenthal DM, Pendall E et al.   Impacts of warming and elevated CO2 on a semi‐arid grassland are non‐additive, shift with precipitation, and reverse over time. Ecol Lett  2016; 19: 956– 66. Google Scholar CrossRef Search ADS PubMed  Naylor D, DeGraaf S, Purdom E et al.   Drought and host selection influence bacterial community dynamics in the grass root microbiome. ISME J  2017; 11: 2691– 704. Google Scholar CrossRef Search ADS PubMed  Nuccio EE, Hodge A, Pett-Ridge J et al.   An arbuscular mycorrhizal fungus significantly modifies the soil bacterial community and nitrogen cycling during litter decomposition. Environ Microbiol  2013; 15: 1870– 81. Google Scholar CrossRef Search ADS PubMed  Nannipieri P, Kandeler E, Ruggiero P. Enzyme activities and microbiological and biochemical processes in soil. In Enzymes in the Environment , New York: Marcel Dekker, 2002, 1– 33. Oliverio AM, Bradford MA, Fierer N. Identifying the microbial taxa that consistently respond to soil warming across time and space. Global Change Biol  2017; 23: 2117– 29. Google Scholar CrossRef Search ADS   Peršoh D. Plant-associated fungal communities in the light of meta’omics. Fungal Divers  2015; 75: 1– 25. Google Scholar CrossRef Search ADS   Philippot L, Raaijmakers JM, Lemanceau P et al.   Going back to the roots: the microbial ecology of the rhizosphere. Nat Rev Micro  2013; 11: 789– 99. Google Scholar CrossRef Search ADS   Price MN, Dehal PS, Arkin AP. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol Biol Evol  2009; 26: 1641– 50. Google Scholar CrossRef Search ADS PubMed  R Core Team, R: A Language and Environment for Statistical Computing . Vienna, Austria: R Foundation for Statistical Computing, 2015, URL https://www.R-project.org/. Ramirez KS, Craine JM, Fierer N. Consistent effects of nitrogen amendments on soil microbial communities and processes across biomes. Global Change Biol  2012; 18: 1918– 27. Google Scholar CrossRef Search ADS   Rillig MC, Mummey DL. Mycorrhizas and soil structure. New Phytol  2006; 171: 41– 53. Google Scholar CrossRef Search ADS PubMed  Romaní AM, Fischer H, Mille-Lindblom C et al.   Interactions of bacteria and fungi on decomposing litter: differential extracellular enzyme activities. Ecology  2006; 87: 2559– 69. Google Scholar CrossRef Search ADS PubMed  Rousk J, Bååth E. Growth of saprotrophic fungi and bacteria in soil. FEMS Microbiol Ecol  2011; 78: 17– 30. Google Scholar CrossRef Search ADS PubMed  Rousk J, Bååth E, Brookes PC et al.   Soil bacterial and fungal communities across a pH gradient in an arable soil. ISME J  2010; 4: 1340– 51. Google Scholar CrossRef Search ADS PubMed  Rudisill MA, Turco RF, Hoagland LA. Fertility practices and rhizosphere effects alter ammonia oxidizer community structure and potential nitrification activity in pepper production soils. Appl Soil Ecol y  2016; 99: 70– 7. Google Scholar CrossRef Search ADS   Saiya-Cork KR, Sinsabaugh RL, Zak DR. The effects of long term nitrogen deposition on extracellular enzyme activity in an Acer saccharum forest soil. Soil Biol Biochem  2002; 34: 1309– 15. Google Scholar CrossRef Search ADS   Santos-Medellín C, Edwards J, Liechty Z et al.   Drought stress results in a compartment-specific restructuring of the rice root-associated microbiomes. mBio  2017; 8: 1– 15. Google Scholar CrossRef Search ADS   Schimel J, Balser TC, Wallenstein M. Microbial stress‐response physiology and its implications for ecosystem function. Ecology  2007; 88: 1386– 94. Google Scholar CrossRef Search ADS PubMed  Six J, Frey SD, Thiet RK et al.   Bacterial and fungal contributions to carbon sequestration in agroecosystems. Soil Sci Soc Am J  2006; 70: 555– 69. Google Scholar CrossRef Search ADS   Steinweg JM, Dukes JS, Wallenstein MD. Modeling the effects of temperature and moisture on soil enzyme activity: Linking laboratory assays to continuous field data. Soil Biology and Biochemistry  2012; 55: 85– 92. Google Scholar CrossRef Search ADS   Steinweg JM, Dukes JS, Paul EA et al.   Microbial responses to multi-factor climate change: effects on soil enzymes. Front Microbiol  2013; 4: 1– 11. Google Scholar CrossRef Search ADS PubMed  Suseela V, Conant RT, Wallenstein MD et al.   Effects of soil moisture on the temperature sensitivity of heterotrophic respiration vary seasonally in an old-field climate change experiment. Global Change Biol  2012; 18: 336– 48. Google Scholar CrossRef Search ADS   Suseela V, Tharayil N, Xing B et al.   Labile compounds in plant litter reduce the sensitivity of decomposition to warming and altered precipitation. New Phytol  2013; 200: 122– 33. Google Scholar CrossRef Search ADS PubMed  Taketani RG, Lançoni MD, Kavamura VN et al.   Dry Season Constrains Bacterial Phylogenetic Diversity in a Semi-Arid Rhizosphere System. Microbial Ecology  2017; 73: 153– 61. Google Scholar CrossRef Search ADS PubMed  Tan X, Chang SX, Kabzems R. Soil compaction and forest floor removal reduced microbial biomass and enzyme activities in a boreal aspen forest soil. Biol Fert Soils  2008; 44: 471– 9. Google Scholar CrossRef Search ADS   Thomson BC, Ostle N, McNamara N et al.   Vegetation affects the relative abundances of dominant soil bacterial taxa and soil respiration rates in an upland grassland soil. Microbial Ecol  2010; 59: 335– 43. Google Scholar CrossRef Search ADS   Tkacz A, Cheema J, Chandra G et al.   Stability and succession of the rhizosphere microbiota depends upon plant type and soil composition. ISME J  2015; 9: 2349– 59. Google Scholar CrossRef Search ADS PubMed  Suttle K, Thomsen MA, Power ME. Species interactions reverse grassland responses to changing climate. Science  2007; 315: 640– 2. Google Scholar CrossRef Search ADS PubMed  Waldrop M, McColl J, Powers R. Effects of forest postharvest management practices on enzyme activities in decomposing litter. Soil Sci Soc Am J  2003; 67: 1250– 6. Google Scholar CrossRef Search ADS   Wallenstein MD, Weintraub MN. Emerging tools for measuring and modeling the in situ activity of soil extracellular enzymes. Soil Biol Biochem  2008; 40: 2098– 106. Google Scholar CrossRef Search ADS   Waring BG, Hawkes CV. Short-term precipitation exclusion alters microbial responses to soil moisture in a wet tropical forest. Microbial Ecol  2015; 69: 843– 54. Google Scholar CrossRef Search ADS   Yang H, Li Y, Wu M et al.   Plant community responses to nitrogen addition and increased precipitation: the importance of water availability and species traits. Global Change Biol  2011; 17: 2936– 44. Google Scholar CrossRef Search ADS   Yergeau E, Bokhorst S, Kang S et al.   Shifts in soil microorganisms in response to warming are consistent across a range of Antarctic environments. ISME J  2012; 6: 692– 702. Google Scholar CrossRef Search ADS PubMed  Zhang K, Shi Y, Jing X et al.   Effects of short-term warming and altered precipitation on soil microbial communities in alpine grassland of the Tibetan plateau. Front Microbiol  2016; 7: 1– 11. Google Scholar PubMed  Zvyagintsev D, Zenova G, Doroshenko E et al.   Actinomycete growth in conditions of low moisture. Biol Bull  2007; 34: 242– 7. Google Scholar CrossRef Search ADS   © FEMS 2017. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

FEMS Microbiology EcologyOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off