Abstract
The soil microbiome’s role in regulating biogeochemical processing is critical to the cycling and storage of soil organic carbon (C). The function of the microbiome under different land management uses has become a focal area of research due to the interest in managing soil C to mitigate climate change. This study investigates the structural and functional response of soil microbiomes from annual monoculture (corn [Zea mays L.]) and perennial diversified (prairie) cropping systems, both under no-till management for bioenergy production. We used a full factorial soil incubation study to understand the influence of temperature and moisture on microbial C decomposition in these soils, with and without addition of cellulose as a model plant residue. Overall, perennial prairie soil supported distinct microbiomes with more diverse prokaryotic and fungal communities compared to annual corn soil. The less diverse corn microbiome was sensitive to the addition of C, resulting in significantly higher respiration compared to prairie, and this increased respiration was amplified under warmer temperatures. In contrast to C loss from the corn soil as carbon dioxide (CO2), prairie soil had significantly higher extracellular enzyme activities and small increases in microbial biomass, illustrating cropping system-specific tradeoffs between microbial C allocation. Specific community structure shifts occurred with added cellulose, where fast-growing, motile decomposers became more abundant under wet conditions, while a small subset of fungi dominated under dry conditions. These differential responses of fungi and bacteria reflect microbial traits important for accessing substrates like plant residues. These changes in community structure due to moisture and cellulose amendment were not necessarily reflected in community function, as potential enzyme activities of most hydrolases were insensitive to temperature and C amendment on this short time scale. Lower respiration occurred in prairie compared to corn soil in response to increased available C and temperature, indicating a more resistant prairie microbiome that may be beneficial when confronted with climate change. These findings support deploying perennial and diversified systems in place of annual monocultures as bioenergy feedstocks, cover crops, buffer strips, or urban greenspaces as part of a land management strategy and highlight the importance of microbial activity in developing sustainable agroecosystems.
Introduction
The role of the soil microbiome in regulating biogeochemical functions has become a major focus of research in microbial ecology because of the interest in managing agricultural soils to mitigate climate change (Schimel and Schaeffer 2012; Jansson and Hofmockel 2020; Yu et al. 2020). A major effect of climate change has been the loss of soil carbon (C) through respiration of carbon dioxide (CO2) (Bond-Lamberty and Thomson 2010; Dacal et al. 2022). However, microbiomes could be leveraged to reduce C loss from soils if we better understood the drivers determining microbial decomposition of plant residues, and in turn how microbial assimilation of plant C is allocated to respiration of CO2 versus biomass C. Therefore, harnessing the soil microbiome to mitigate the negative effects of climate change has become pressing in recent years.
The ability of soil microorganisms to decompose plant inputs to create organic matter and release nutrients is fundamental to creating more productive and sustainable crop production systems, especially in the context of bioenergy feedstocks, which may offset fossil fuel use and concurrently accrue soil C (National Academies of Sciences, Engineering 2019). For example, perennial grasses are of particular interest for bioenergy production because they require lower inputs of fertilizers and other chemicals, mitigate nitrogen (N) runoff into surface and groundwater, and reduce management costs compared to conventional crops such as corn (Zea mays L.) (Liebman et al. 2013). Biomass yields of perennial grasses can be as high or higher than corn (Sanford et al. 2016; Yang et al. 2019). Comparatively, perennial grass systems have deeper roots with higher biomass and increased rhizodeposition, resulting in increased C inputs into the soil (Jarchow et al. 2015; McGowan et al. 2019) relative to annual crops. Increased C inputs may be linked to long-term accrual and stabilization of soil organic matter (SOM) (Sanford et al. 2022). These benefits of perennial crops are amplified in diversified agroecosystems compared to monocultures (Werling et al. 2014; Jesus et al. 2016). For instance, Landis et al. (2018) found increased abundance of soil methane consumers in diverse versus monoculture perennial systems, and higher levels of other ecosystem services such as diversity of pollinating insects and pest suppression. Conversion of bioenergy crops to mixed native perennials is also associated with greater C inputs into the soil and reduced N runoff (Liebman et al. 2013).
The benefits of diversified cropping systems, like prairie, extend belowground and have been shown to select for a more diverse and functionally resilient microbiome (Zak et al. 2003; Eisenhauer et al. 2010; Milcu et al. 2013) that may be better able to withstand the stress of climate change and variability. Plant C inputs to the soil via exudation, litter, and root detritus differ across plant species in magnitude and biochemical composition (Herz et al. 2018). Such differences can drive changes in microbial biomass (Liang et al. 2012, 2016; Thakur et al. 2015; Mackelprang et al. 2018) and community composition (Jesus et al. 2016; Oates et al. 2016; Mackelprang et al. 2018), as well as microbial function as measured by potential extracellular enzyme activity (EEA) (Bach and Hofmockel 2015; Upton et al. 2020) and respiration rates (Tufekcioglu et al. 2001; von Haden et al. 2019). For example, our previous work revealed that a restored native prairie had significantly richer and more diverse microbiomes compared to conventional corn (Bach et al. 2018; Upton et al. 2018). Selection for a diverse microbiome is likely due to not only plant diversity, but also perennialism (Robertson et al. 2017). Like plant diversity, microbial diversity is important for maintaining ecosystem functions. Previous studies showed that artificially reducing soil microbiome richness slows litter composition (Wagg et al. 2019) and increases the preference for cycling of highly labile C sources (Maron et al. 2018)—phenomena which would dampen the soil’s effectiveness as a C sink. Highly diverse bacterial and fungal communities are also linked to resilience against warming induced remineralization of labile pools of SOM to CO2 (Xu et al. 2021). Soil microbiome diversity influences additional ecosystem functions such as turnover of complex SOM (Eisenhauer et al. 2010; Xu et al. 2021), impacting C storage and N availability, which may enhance plant productivity (Lau and Lennon 2011; Schnitzer et al. 2011).
Microbial diversity and function respond to environmental factors such as moisture and temperature, which are currently shifting with our changing climate. Expected temperature increases during summer are already observable in the Midwest (Liu and Basso 2020), which is a key agricultural region where production of bioenergy crops has expanded to meet demands for cellulosic ethanol production. Warming has been found to impact microbial community structure with increases in saprotrophic fungi found in a meta-analysis (Gang 2019), increases in overall bacterial to fungal ratios (Zhang et al. 2005), and selection for fast-growers (Docherty and Gutknecht 2019). Despite repeatable changes in community structure, studies on the impact of increased temperature on microbial biomass have been less consistent. For example, studies measuring temperature effects on total microbial biomass are mixed and have found that higher temperatures reduce biomass (Zhang et al. 2013; Suzuki et al. 2016; Gao and Yan 2019), have no effect (Zhou et al. 2013a), or increase microbial biomass (Bell et al. 2010; Xu and Yuan 2017). The impact of temperature on microbial extracellular enzymes, the direct agents of decomposition, is also inconsistent. A meta-analysis found an overall positive correlation between EEA and temperature, but with high variation across the included studies (Meng et al. 2020). Bell et al. (2010), Steinweg et al. (2013), and Thakur et al. (2019) found no effect of increased temperatures alone on EEA across pasture, abandoned agricultural field, and temperate grassland systems. However, Hou et al. (2016) found significant increases in a no-till crop rotation system, and Zhou et al. (2013b) found a mix of both increases and decreases on different enzymes that varied with soil depth under elevated temperatures. It remains challenging to disentangle the myriad factors (such as soil type, crop system, season, or historical precipitation regime) that may be contributing to observed variation in measured responses of microbial biomass and EEA to increased temperature, making these microbiome features difficult to predict.
Modeling suggests that some of the variation of microbially mediated C-cycling with increased temperature may be attributed to alterations in moisture interacting with temperature (Xu et al. 2015). In addition to overall temperature increases, elevated atmospheric CO2 concentrations and regional temperature increases are altering rainfall patterns across the Midwest region (Parmesan and Yohe 2003; Angel et al. 2018). Much of the Midwest and Great Plains region is experiencing an increase in severe droughts as well as extreme rainfall events (Pryor et al. 2014; Shafer et al. 2014; Ford et al. 2021; Grady et al. 2021). At a microbially relevant scale, soil habitats become fragmented under drought conditions, with films of water limited to small pores spaces and air spaces preventing dispersal of bacteria and access to substrates (Dechesne et al. 2008; Vos et al. 2013; Wolf et al. 2013). This fragmentation can promote biodiversity by creating habitat niches and reducing competition. Reduced substrate access can alter C degradation (Ebrahimi and Or 2016) and may contribute to observed temporary C accumulation during drought conditions (Canarini et al. 2017; Schaeffer et al. 2017; Karlowsky et al. 2018). In contrast, under high moisture conditions, connected pores favor bacteria that are motile and can easily access labile substrates (Bérard et al. 2015). As a result, fungal communities in particular may be less easily impacted by drought than bacterial communities (Yuste et al. 2011; Manzoni et al. 2012; de Vries et al. 2018; Glassman et al. 2018) because their hyphae can bridge dry spaces in the soil and cycles of drying-rewetting influence bacterial growth to a greater extent than fungal growth (Bapiri et al. 2010). Drought also induces warming, which generally increases microbial decomposition of roots (Zhou et al. 2022). The decomposition of root biomass can be slowed under low moisture conditions (Liu et al. 2021), but interacts with temperature to alter decomposition. Therefore, managing soil C depends on unraveling the response of the soil microbiome to multiple interacting factors of climate change.
Here we aimed to understand how microbiomes selected by annual versus perennial bioenergy crops differ in their decomposition activity under simulated climate change in the form of altered temperature and moisture. We chose cellulose as a model substrate for plant inputs and included two soil aggregate sizes that differ in their C content. We tested the following questions:
How do diversified perennial versus monoculture annual no-till cropping systems affect soil microbiome diversity, structure, and decomposition function measured as enzyme production (EEA), respiration (CO2), and growth (microbial biomass C)?
How do moisture and temperature interact to alter soil microbiomes and their functions?
Our central hypothesis was that the microbiome selected by cultivation of soil with a diverse perennial plant community would also be more diverse and resistant to perturbation than a microbiome from an annual monoculture crop. As such, we expected the magnitude of response to temperature and moisture changes to be smaller for prairie soil than corn. We expected prairie microbiomes to maintain the majority of their microbial diversity and structure in the face of perturbation by simulated climate change, and corn microbiomes to lose diversity and shift community structure. We also expected prairie microbiomes to be primed for using C more efficiently, by allocating more C to increased biomass and extracellular enzyme production, relative to respiration, compared to corn microbiomes in the face of increased temperature and altered soil moisture.
Materials and Methods
Site Description. Soil in the experiment was from the ongoing Iowa State University Comparison of Biofuel Systems (COBS) experimental site located on the South Reynoldson Farm in Boone County, Iowa (41°55′14.42″ N, 93°44′58.96″ W), which was initiated in 2008 (see Jarchow and Liebman [2013] for a detailed site description) by tilling the site and planting the cropping treatments. Soils consist mainly of Webster silty clay loam (fine-loamy, mixed, superactive, mesic Typic Endoaquolls) and Nicollet loam (fine-loamy, mixed, superactive, mesic Aquic Hapludolls) with a mean pH of 6.7. Prior to establishment of the site the area was planted with corn and soybean (Glycine max [L.] Merr.) rotations. The present study includes two of the experimental bioenergy cropping systems: no-till continuous corn, an annual monoculture left fallow in winter, and fertilized planted tallgrass prairie (31 native perennial species). Nitrogen is supplied as urea-ammonium nitrate (NO3−) in early spring, but corn plots are fertilized approximately twice as much as the fertilized prairie plots (approximately 178 kg N ha−1, depending on spring NO3− tests, versus 84 kg N ha−1 [Daigh et al. 2015]). At harvest, ~50% of the corn stover and ~75% of the prairie residue is harvested and removed from the field plots (Upton et al. 2019). Four replicate blocks contain one 27 m × 61 m plot of each planting treatment in a randomized complete block design (total n = 8). Soils in the area are rainfed and poorly drained, therefore a subsurface tile drainage system is installed at the site (Daigh et al. 2015). Fertilized prairie soil displays higher aggregation (3.62 ± 0.12 versus 3.26 ± 0.14 mean weighted diameter [mm]) and lower bulk density (1.42 ± 0.02 versus 1.51 ± 0.02 g cm−3) than corn soil. Additional soil physical properties in each management system are reported in Bach and Hofmockel (2016).
Soil Collection. During early season growth (June 8, 2014), five soil cores were taken (7.5 cm diameter × 10 cm deep) from each plot and transported back to the laboratory on ice. The soil cores were gently broken though 8 mm sieves and subsequently fractionated into aggregate size classes: micro-aggregates (<250 µm), small macroaggregates (250 µm to 1 mm), medium macroaggregates (1 to 2 mm), and large macroaggregates (>2 mm) by the optimal moisture approach as described by Bach and Hofmockel (2014). The respective aggregate size fractions from the four replicate plots of each cropping system were pooled and stored at 4ºC until used in the microcosm incubations. The water holding capacity (WHC) of the small macroaggregates and large macroaggregates were determined by filling 10 g of soil into cylinders at an approximate bulk density of 1 (Margesin and Schinner 2005). Cotton fabric secured with a rubber band was used to secure the bottom of the cylinder. Cylinders were saturated for 24 h then drained until the weight was constant. Subsequently, the gravimetric water content was determined.
Experimental Design. Prairie and corn microcosm incubations were inoculated with small or large aggregate size fractions. In total, a 2 × 2 × 2 × 2 × 2 factorial experiment was established corresponding to two bioenergy crop systems, two aggregate size classes, two levels of C availability (labile C amended or unamended controls), two moisture levels (25% [“dry”, simulating drought conditions] or 75% [“wet”, simulating postrainfall conditions] of total WHC of the respective aggregate size class), and two temperatures (15ºC or 25ºC) (figure 1). Three replicates were established per treatment resulting in total 96 microcosms. Due to limited mass of small macroaggregates obtained from the corn ecosystem, 5 of the 16 treatments were reduced to duplicate microcosms.
Microcosms were established by adding 10 g of soil of the respective aggregate size to sterile 100 mL serological glass vials. The microcosms were preincubated at 20°C for seven days at optimal moisture, after which 20 mg cellulose g−1 soil as avicel (Sigma-Aldrich, Inc, St. Louis, Missouri) was added to the cellulose treatment (“labile C amended”) microcosms and all treatments were adjusted to the respective moistures. Microcosms were closed with butyl stoppers and aluminum caps and incubated at the respective treatment temperatures for 10 days.
Soil Respiration. Microbial metabolism was monitored via respiration rates by measuring the CO2 cumulative concentration in the headspace of the microcosms with a LI-7000 CO2/water (H2O) Gas Analyzer (LI-COR Biosciences Inc, Lincoln, Nebraska). At day 0, 1, 4, 7, and 10 of the experiment, 200 µL was removed from the microcosm headspace and injected into the sensor with a Pressure Lok syringe (Valco Instruments, Houston, Texas). Headspace was replaced with CO2-free air to maintain standard air pressure. Carbon dioxide concentrations were calculated using a reference gas (1,000 ppm) adjusting for dilution with replacement air of the headspace over time. To be sure that the microcosms did not go anaerobic, we measured oxygen (O2) concentrations in the headspace over time with an O2 analyzer (Quantek, Grafton, Massachusetts). In addition to calculating CO2 respired per gram dry soil (cumulative CO2), we also normalized cumulative CO2 per unit of soil microbial biomass (see below) to analyze biomass specific respiration.
Carbon and Nitrogen Pools. Microbial biomass and extractable C and N were measured in 0.5 M potassium sulfate (K2SO4) soil extracts of fumigated and nonfumigated soil, respectively. Extractions were performed on 5 g soil as described by (Brookes et al. 1985). Extracts were analyzed for nonpurgeable organic C and total N via combustion catalytic oxidation (Shimadzu TOC-L analyzer, Shimadzu Corporation, Columbia, Maryland), and conversion factors of 0.45 for C and 0.54 for N were used to convert organic C and N to microbial biomass (Brookes et al. 1985; Vance et al. 1987).
DNA Extraction and Amplicon Sequencing. DNA was extracted from soil using the PowerSoil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, California) according to the manufacturer’s protocol. DNA extracts were quantified by nanodrop (Nanodrop ND-1000, Thermo Fisher Scientific, Waltham, Massachusetts) and stored at −80°C. 16S and ITS amplification, library preparation, and sequencing were performed by Argonne National Lab NGS core facility following Earth Microbiome protocols (Caporaso et al. 2011). The prokaryotic 16S ribosomal rRNA gene from both archaeal and bacterial V4 region was amplified with primer pair 515F/806R (Apprill et al. 2015; Parada et al. 2016) and the fungal ribosomal ITS1 region amplified with primer pair ITS1F/ITS2 (White et al. 1990). Sequencing was performed on an Illumina MiSeq with 151 × 151 (16S) or 251 × 251 (ITS) pair-ended Illumina chemistry.
Sequences for both bacteria and fungi were processed using the open source pipeline “hundo” (https://github.com/pnnl/hundo), which implements methods described in Caporaso et al. (2010). Briefly, raw reads are demultiplexed using EA-Utils (Aronesty 2013) with zero mismatches allowed in the barcode sequence. Adaptor and PhiX sequences were removed and quality filtered by BBDuk2 (Bushnell 2014) with matching k-mer length of 31 bp at a hamming distance of 1. Reads shorter than 51 bp were discarded. USEARCH (Edgar and Flyvbjerg 2015) was used to merge reads with a minimum length threshold of 175 bp and maximum error rate of 1%. Dereplicated sequences (minimum sequence abundance of two) were clustered using the distance-based, greedy clustering method of USEARCH (Edgar and Flyvbjerg 2015) at 97% pairwise sequence identity among operational taxonomic unit (OTU) member sequences. Chimeric sequences were removed by de novo prediction during clustering. Prokaryotic taxonomy was assigned to 16S OTU sequences using BLAST (Camacho et al. 2009) alignments followed by least common ancestor assignments across SILVA database version 123 clustered at 99% (Camacho et al. 2009; Quast et al. 2013). OTU seed sequences were filtered against SILVA database version 123 clustered at 99% to identify chimeric sequences. Fungal taxonomy was assigned to ITS OTU sequences using BLAST alignments followed by least common ancestor assignments across Unite version 7 database trained on the ITS2 region (Kõljalg et al. 2013), and, to identify chimeric OTUs, the OTU seed sequences were filtered against Unite UCHIME reference database version 01.01.2016 (Kõljalg et al. 2013) using USEARCH. Read counts were normalized by cumulative sum scaling (CSS) normalization (Paulson et al. 2013) implemented in R using the package phyloseq (McMurdie and Holmes 2013).
Potential Extracellular Enzyme Activity. The potential activities of several extracellular enzymes were measured in soil collected from microcosms after the 10-day incubation as described by DeForest (2009) with modifications described by Bach and Hofmockel (2014). This included the C-cycling enzymes β-glucosidase (BG), β-xylosidase (BX), and cellobiohydrolase (CB), the N-cycling enzyme N-acetyl-glucosaminidase, P-cycling enzyme acid phosphatase (AP) using 4-Methylumbelliferyl- linked substrates, and the peptidase leucine-aminopeptidase (Leu) using a 4-methylcoumarin-linked substrate. Analyses were performed on 1 g frozen soil sample in 125 mL of 100 mM Tris-Maleate buffer adjusted to pH 6.7. Sample fluorescence (360 nm excitation and 460 nm emission) was read using a microplate reader (BioTek, Winooski, Vermont) after 2 h incubation. Enzymatic activity was calculated as nanomoles per hour per gram of dry soil (nmol h−1 g−1) according to DeForest (2009), German et al. (2011), and German et al. (2012). The measured enzyme actives are regarded as potential activities as the assays are conducted with saturating substrate concentration that do not reflect in situ conditions.
Statistical Analysis. All analyses were run in R (R Core Team 2021). Analysis of variance (ANOVA) was used to examine variation in cumulative CO2 respired, microbial biomass, and potential enzyme activities across treatments after the 10-day incubation. A series of two-way ANOVA models, allowing the fixed effects to interact, were used to examine the effects of crop × aggregate size × moisture × temperature × C addition. Tukey’s honestly significant difference (HSD) post-hoc analysis was used to disentangle differences among the various interaction levels. Microbial community alpha diversity indices such as richness, Shannon’s evenness, and Shannon’s diversity index were calculated from CSS normalized prokaryotic and fungal OTU read counts in the package vegan. Microbial beta diversity patterns across treatments were assessed with perMANOVA and visualized using NMDS. perMANOVA were run in R using the adonis function in package vegan (Oksanen et al. 2017), with 999 permutations and Bray-Curtis dissimilarity indices. Continuous variables (cumulative CO2, enzyme activities, and C and N values) were fit as vectors onto the ordination coordinates using envfit from the package vegan. Despite the differing C contents, no major differences in treatment effects were found between the communities or extracellular enzymes in the two aggregate groups. Therefore, aggregate size classes were combined for analysis to reduce complexity.
Results and Discussion
Microbiome Responses. Prairie microbiomes were more diverse than corn microbiomes, particularly in the fungal components. Dry conditions amended with labile C dramatically altered fungal communities, while labile C added under wet conditions dramatically decreased prokaryotic diversity and altered prokaryotic community structure.
Prokaryotic Community Structure and Diversity by Treatment. Prairie cropping system selected for a distinct and more diverse microbial community than corn. Moisture had a clear impact on prokaryotic communities involved in cellulose decomposition, as demonstrated by multivariate analysis. With water and cellulose amendments, both corn and prairie microbiomes shifted away from microbiomes of all other treatments (wet, dry, dry + cellulose) (figure 2), yet the two crop microbiomes remained distinct from each other. The separation of the corn and prairie communities was significantly correlated with potential enzyme activity (figure S1a), suggesting the differences in community structure correspond with a subset of microbes that are able to access and decompose the cellulose, including cross-feeders that can assimilate products of cellulose decomposition. These differences are supported by significantly greater enzyme activity measurements in prairie compared to corn soils, suggesting that decomposition capacity differentiated the two microbiomes. The same axis was negatively correlated with biomass specific respiration, demonstrating the tradeoff between microbial C allocation to enzyme production versus loss from the soil as CO2.
Respiration increased by three-fold more when labile C was added to corn versus prairie soils, increasing by ~110 µg CO2 g−1 soil in corn versus ~37 µg CO2 g−1 soil in prairie relative to unamended soils. These observations reveal that prairie microbiomes were functionally superior in terms of decomposition potential, while corn microbiomes had greater C loss due to higher cumulative rates of microbial respiration. Together, our results confirm that diversified perennial cropping systems support more diverse soil microbiomes with enhanced ability to decompose plant residues. Further, we find that loss of C by microbial respiration is reduced in diversified systems. Increased decomposition coupled with diminished C loss via microbial respiration supports the potential for diversified cropping systems to help mitigate greenhouse gas emissions from agroecosystems (Szymanski et al. 2019; Sanford et al. 2022).
The decomposition response under wet conditions was accompanied by a change in prokaryotic community structure (figure 2, figure S1, and table S1). This was true for both crop microbiomes and was especially pronounced in soils derived from corn plots. In the presence of cellulose amendments, prokaryotic diversity decreased dramatically (figure 3 and table 1) under wet conditions. This is in line with the understanding that fast growers often respond rapidly to labile C substrates, outcompeting slower growing bacteria and fungi (de Graaff et al. 2010; Goldfarb et al. 2011; Hicks et al. 2022). Because corn soils support lower microbial diversity, this response was substantially more pronounced in corn microbiomes where fewer competitors enabled the dominance of a select subcommunity of bacteria, particularly motile members of Flavobacterium, which are known cellulose degraders (Ruff 2022). Wet conditions that encourage access to substrates via high pore connectivity (Carson et al. 2010) give motile and fast-growing species a competitive advantage for resource acquisition. Reductions in diversity and increased access to resources have been linked to higher decomposition rates of soil organic matter (Tóth et al. 2017), similar to the findings in this study.
Temperature was a much weaker driver of microbiome responses than soil moisture (table 1). The decreased diversity under wet conditions with cellulose amendments was exacerbated by incubation at cold temperatures (15°C). Diversity loss was associated with moderate increases in the relative abundance of only a few taxa, along with simultaneous decreases in dozens of other taxa under cellulose addition. These results support our hypothesis that diverse perennial systems drive higher microbial biodiversity and is in agreement with previous work (Bach et al. 2018; Upton et al. 2019), although some work has found conflicting results (Mackelprang et al. 2018).
Fungal Community Structure and Diversity by Treatment. Consistent with prokaryotic responses, the perennial prairie ecosystems supported distinct fungal diversity (figure 4) and enhanced function compared to corn systems. Elevated fungal diversity was consistent across treatments (moisture, temperature, and C addition) (figure 5 and table 1), supporting the enhanced resistance of soil microbiomes under diversified plant cover (De Vries et al. 2012; Milcu et al. 2013). In contrast to the prokaryotic response, we found that fungal communities were differentiated in the presence of cellulose additions under dry, not wet conditions (figure 4 and figure S2). This is likely because fungi prefer aerobic conditions and are well adapted to access labile substrates via mycelial growth. The ability of fungal mycelia to bridge air spaces has been shown to provide a competitive advantage for substrate acquisition in dry compared to wet conditions (Guhr et al. 2015; Treseder et al. 2018; Bhattacharjee et al. 2020). The diminished diversity was amplified in corn compared to prairie microbiomes and once again this effect was enhanced by incubation at lower temperatures (15°C). Although loss of diversity occurred in different moisture treatments for prokaryotes and fungi (wet + labile C versus dry + labile C, respectively), community alteration was similar in that it supported large increases in a small number of taxa and moderate decreases in many taxa. The changes in fungal community structure demonstrate that a select subset of the fungal community outcompetes other taxa to decompose cellulose. A previous incubation leveraging stable isotope tracers at moderate moisture levels also found a small subset of fungi were responsible for the decomposition of cellulose (Koechli et al. 2019), including members of the class Sordariomycetes, which also increased in our study. Dry conditions likely favor these filamentous taxa that can bridge between discrete spatially disconnected pockets of resources, while under saturated conditions they lose this advantage (Yuste et al. 2011; Worrich et al. 2017; Bhattacharjee et al. 2020). This is further supported by previous work showing that lower moisture appears to reduce competition in fungal communities (Hawkes et al. 2011). Fungal activity can dominate respiration and decomposition of SOM in dry soils, but increased temperature reduced this trend (Yuste et al. 2011). Similarly, fungal growth was reduced in agricultural soils at higher temperatures (PietikÃ¥inen et al. 2005) and warming appeared to stress fungal communities, reducing fungal-driven decomposition in prairie soil (Guo et al. 2018; Docherty and Gutknecht 2019).
Moisture is a controlling factor of habitat connectivity in soils, limiting microbial access to substrates and diffusion of extracellular enzymes involved in cellulose decomposition. Our dry treatments were equivalent to drought conditions in our soil, while the wet condition modeled soil moisture levels following a rainfall. Together the shifts observed in the soil microbiome in response to water availability demonstrate that the effects of moisture on fungal diversity are distinct from prokaryotic diversity, where fungal decomposers have an advantage under dry conditions and prokaryotic decomposers under wet. These differences have important implications for microbial function as climate shifts occur in mean and extreme temperature and precipitation regimes (Jansson and Hofmockel 2020). Our results indicate that soil desiccation may favor fungal taxa that are able to access substrates, while extreme rainfall events will impact prokaryotic decomposers to a greater extent, and that for both these stressor events will be stronger if they take place in cooler months. The strong interaction of moisture with C amendment indicates that access to the C due to habitat connectivity and diffusion of nutrients is more important than C quality and quantity in the soil for C partitioning, given the lack of change in extractable C or N due to moisture by itself (table 2).
Effect of Cropping System on Soil Respiration. The availability of a labile C source under simulated climate change caused dramatically increased soil respiration by the corn microbiome. Soil respiration regulates soil C storage as the dominant pathway of C loss from ecosystems (von Haden et al. 2019), and therefore alteration of the flux of C inputs into biomass and SOM versus CO2 is a key metric in soil health. In the absence of added labile C, cumulative CO2 g−1 soil was slightly higher in prairie soils than in corn (F = 96.26, p < 0.0001), and cumulative CO2 was higher at 25°C compared to 15°C (F = 205.45, p < 0.0001; figure 6a). However, the addition of cellulose, a labile C source simulating fresh plant residue, significantly increased soil respiration, especially for the corn microbiomes (F = 391.17, p < 0.0001). The respiration response of corn microbiomes to cellulose inputs was 6-fold higher than unamended soil, while the prairie microbiome response was only 2-fold higher than unamended conditions. This effect was amplified under warmer temperatures, as evidenced by the highest cumulative respiration in C-amended 25°C treatments. Cellulose addition did not significantly increase microbial biomass C (MBC) in any of the treatments (F = 0.11, p = 0.7414; table 2 and figure S3), indicating that additional cellulose amendment resulted mostly in increased enzyme production or microbial metabolism and maintenance, but not significant net growth. Normalization of respiration to MBC revealed that respiration per unit biomass (or biomass-specific respiration rates, mg CO2 mg−1 MBC) was still significantly increased by added C (F = 29.33, p < 0.001), but there was a strong C amendment by cropping system interaction, such that biomass-specific respiration of prairie microbiomes only increased marginally, while corn microbiomes increased on average by 5-fold (figure 6b). The differences in corn compared to prairie microbiome responses were most pronounced in the 25°C dry treatment, where corn microbiomes generated a 15-fold increase in respiration. These results indicate that some of the respiration increase in prairie soils was accounted for by small, but not statistically significant, differences in microbial biomass C; whereas, in corn samples, the increase in respiration was not accounted for by concurrent changes in biomass and has the potential to produce large fluxes of greenhouse gases from the soil to the atmosphere. These results support our hypothesis that prairie microbiomes use C more efficiently than corn microbiomes, resulting in multiple benefits to C cycling and soil fertility. Cumulative CO2 was higher under wet conditions for most combinations of treatments, but this effect was small compared to the influences of crop and temperature. The biomass-specific respiration rates also reduced differences between dry versus wet conditions in prairie soil, suggesting that microbial C use efficiency was similar under both moisture conditions, and that treatment differences in CO2 respired per gram soil are due to small differences in microbial biomass. Our results indicated that moisture alone is not the primary driver of higher respiration in these soils; instead, the microbiome is strongly C limited, albeit to a lesser extent in prairie soils. Investment of labile C into biomass production or maintenance by the more diverse prairie microbiome represents an initial step in stabilizing C as SOM.
Carbon and Nitrogen Pools. Differences in respiration between corn and prairie soils did not appear to be caused by differences in resource quantity, as there was no statistical difference in extractable organic C between cropping systems, even 10 days after C addition (table 2 and figure S4a). However, there was more extractable N in corn than prairie at the end of the incubation (F = 5.19, p = 0.0257), and extractable N was lower in C-amended soil after 10 days (F = 13.13, p = 0.005; figure S4b), reflected in a shift in the extractable C/N ratio. These results suggest that C addition primed decomposition of the native SOM (Liu et al. 2020) and altered N pools. Microbial biomass N (MBN) was also lower in C-amended treatments (F = 10.66, p = 0.0017), but this decrease was not large enough to significantly alter the C/N in microbial biomass. Inputs of C into soil from corn are less than in prairie systems (Jarchow et al. 2015), due to both lower root biomass and rhizodeposition. The corn plots at our study site also receive twice as much N fertilization as prairie (corn = 178 kg N ha−1; prairie = 84 kg N ha−1), causing the corn microbiome to be more C limited as a result of reduced plant inputs and higher N inputs. The addition of labile C consequently stimulated C use in continuous corn soil and increased respiration rates. Although other work has found reductions in microbial respiration with N fertilization (Ramirez et al. 2012), it appears the higher N fertilization in our corn system is not enough to compensate for acclimation of the microbiome to differences in corn and prairie C substrate quality.
Effects of Cropping System, Moisture, and Temperature on Community Function. Potential enzyme activities were systematically higher in prairie compared to corn soil, but contrary to our predictions, temperature had no discernable effect on enzyme activities, and the activity of C-cycling enzymes was increased under dry conditions. The diverse prairie microbiomes supported greater potential enzyme activity (µmol MUB g−1 soil h−1 for β-glucosidase (BG), β-xylosidase (BX), and cellobiohydrolase (CB); p < 0.0001; figure 7 and table S2), which are all hydrolases involved in decomposition of cellulose and related oligosaccharides. Prairie microbiomes also produced greater enzyme pools to decompose substrates rich in C and N (figure S5). This includes N-acetyl-glucosaminidase, which targets chitin decomposition products (p < 0.005). Potential activity rates were higher under dry conditions, suggesting these may be of fungal origin. Prairie microbiomes also supported greater pools of leucine aminopeptidase compared to corn microbiomes (p < 0.0001). Leucine aminopeptidase targets peptide and protein residues that are rich in organic N. These results demonstrate that the prairie microbiome has superior enzymatic capability compared to corn, enabling acquisition of both C and N from plant residues and labile SOM pools (Allison and Jastrow 2006; Mahaney 2010; Bach and Hofmockel 2015, 2016; Upton et al. 2018). Both prairie and corn microbiomes responded to cellulose additions with small but significant increases in specific rates of enzyme activity for β-glucosidase and β-xlylosidase activities (p < 0.05). Interestingly, on average across the two microbiomes, the addition of labile C in the form of cellulose did not alter the activity of cellobiohydrolase, the enzyme that cleaves internal cellulose bonds, on either a per gram soil basis or when normalized to MBC, at least on this short time scale. This could be in part because the pool of cellobiohydrolase enzymes within the soil were adequate to meet C demands and there was no C limitation with the cellulose addition. However, when moisture was elevated, cellobiohydrolase activities decreased to around half the activity detected under dry conditions, providing further evidence that the fungal communities were dominating decomposition activities under dry conditions (Liu et al. 2022).
We observed no increases in any C-degrading enzymes under increased temperature, even in the presence of added C substrate. Cellulase activity may be modulated by substrate stoichiometry and cellulase production limited by N and phosphorus (P) availability (Tian and Shi 2014; Bach and Hofmockel 2015; Jian et al. 2016). The lack of strong or consistent patterns in enzymatic potential has been observed repeatedly by other studies. Docherty and Gutknecht (2019) detected little change in overall potential enzyme activities in response to cellulose amendment at lower temperature. A meta-analysis found high variability in enzyme activity with only a small increase with warming (Meng et al. 2020). Likewise, neither soil moisture nor warming caused an appreciable increase in enzyme activities in a field of mixed grasses and forbs (Steinweg et al. 2013). Chen et al. (2020) invoked a drawdown of available substrates for cellulases as a possible explanation for not detecting increase in cellulase activities with temperature, although ligninase was responsive to temperature. Our results therefore align with previous research and point to lack of dramatic shift in microbial extracellular enzyme production induced by temperature, even with substrate additions. Alternatively, small changes could be masked due to slow turnover and degradation of enzymes in mineral soils, or due to the stabilization of extracellular enzymes on soil mineral surfaces, which can interfere with detectability of small shifts in potential activity in these assays (Blankinship et al. 2014). Because fluorescent enzyme assays are measured under unlimited substrate concentrations, and eliminate diffusion limitation on decomposition rates, they can measure activity of enzyme that may have been immobilized and inactive in situ, and therefore underestimate differences in in situ enzyme activity that may be present with variable moisture regimes or substrate access.
Implications for Land Use Management. We found a distinct and more diverse soil microbiome under prairie cultivation as compared to an annual monoculture of corn. More importantly, the prairie microbiome appears to be more resistant to climate disturbances than the corn microbiome, with smaller increases in respiration in response to increased temperature or moisture. Decomposition of added labile C caused a striking alteration of respiration, particularly in dry and warm corn microbiomes. The conditions mimic a hot summer drought, which are expected to increase in frequency with climate change. These results are therefore an indication that C inputs, as from rhizodeposition or litter inputs, are less likely to be incorporated into SOM and stabilized under a corn microbiome. The large increase in respiration when normalized to microbial biomass, which reflects C use efficiency, indicated that warm and dry conditions are causing an increase in maintenance respiration, without paired increases in microbial biomass. Microbial biomass and growth are important for generating SOM, through stabilization of microbial residues or necromass. Therefore, reduced biomass production may be damping the microbial processing of plant derived inputs into SOM and concurrently shunting C out of the soil via respiration. Our results support the increased use of diverse perennial systems for bioenergy crop production, riparian buffer strips, and urban spaces as a climate mitigation strategy.
Summary and Conclusions
The response of soil microbiomes to changing environmental conditions is a critical aspect of agroecology, with a promising role in climate mitigation. Yet, the interactive effects of differential precipitation and warming on decomposition dynamics have been challenging to resolve due to the complexity of the soil microbiome and decomposable substrates in the soil environment. Here we accounted for multiple interactive effects to examine how annual monocultures and diversified perennial cropping systems select for distinct soil microbiomes with different decomposition profiles and contrasting responses to changes in moisture and temperature. Cropping system had a greater effect on soil respiration than moisture, temperature, or C input in our study. The impact of C amendment was highly contingent on a complex interplay of the soil microbiome, temperature, and moisture conditions. Soil microbiomes from a monoculture of annual corn produced more CO2 than the diverse perennial prairie microbiomes when C was added, and this effect was amplified at a higher temperature, indicating these microbiomes are sensitive to changes in C inputs and climate conditions. These results support that extensive perennial root systems of diversified prairies select for microbiomes able to efficiently use labile C, reducing loss of soil C as CO2, and potentially enhancing C storage compared to corn-selected microbiomes. The limited impact of C addition and warmer temperatures, combined with lower respiration in prairie compared to corn soil, points to a more resilient microbiome in the face of climate change. The response of microbiomes to climate alteration in diverse perennial crops should be an area of further research to delineate the impacts in other systems. This study highlights the potential for land management decisions as a tool for climate mitigation, especially when considering bioenergy feedstocks, cover crops, or diversifying rotations.
Supplemental Material
The supplementary material for this article is available in the online journal at https://doi.org/10.2489/jswc.2023.00069.
Acknowledgements
We thank Iowa State University’s College of Agriculture and Life Sciences, Michael Thompson, and Robert Horton for supporting the field experiment; Alexandra Wolf for input on the experimental design; and graduate assistants Rachel Upton and Elizabeth Bach and undergraduate assistants Montana Smith, Christina Davis, Jacinta Misra, and Jack Nielsen for their assistance with conducting field sampling and soil processing. This program is supported by the US Department of Energy (DOE), Office of Science, through the Genomic Science Program, Office of Biological and Environmental Research, under (award number DESC0010775) and FWP 70880. A portion of this work was performed in the William R. Wiley Environmental Molecular Sciences Laboratory (EMSL), a national scientific user facility sponsored by Office of Biological and Environmental Research and located at the Pacific Northwest National Laboratory (PNNL). PNNL is a multiprogram national laboratory operated by Battelle for the DOE under Contract DE-AC05-76RLO 1830.
- Received May 3, 2022.
- Revision received November 21, 2022.
- Accepted November 28, 2022.
- © 2023 by the Soil and Water Conservation Society