Plant Soil (2011) 339:401–412 DOI 10.1007/s11104-010-0592-y
REGULAR ARTICLE
Soil microbial community function, structure, and glomalin in response to tall fescue endophyte infection Jeffrey S. Buyer & David A. Zuberer & Kristine A. Nichols & Alan J. Franzluebbers
Received: 9 July 2010 / Accepted: 24 September 2010 / Published online: 16 October 2010 # Springer Science+Business Media B.V. (outside the USA) 2010
Abstract Tall fescue [Lolium arundinaceum (Schreb.) S.J. Darbyshire] is naturally infected with a fungal endophyte, Neotyphodium coenophialum, which produces toxic ergot alkaloids that negatively affect herbivores and may alter soil microbial communities. A 60-week mesocosm study with a factorial arrangement of soil type (clay loam and loamy sand) and endophyte infection (with and without) was conducted to determine changes in soil microbial community function (substrate utilization using Biolog), structure (phospholipid fatty acid profile), and glomalin concen-
Responsible Editor: Eric Paterson. J. S. Buyer USDA—Agricultural Research Service, Building 001, Room 140, BARC-West, Beltsville, MD 20705, USA D. A. Zuberer Department of Soil and Crop Sciences, Texas A&M University, College Station, TX 77843, USA
tration. Microbial utilization of carbohydrate, carboxylic acid, and miscellaneous substrate groups was lower in soil planted to endophyte-infected tall fescue than in soil planted to endophyte-free tall fescue. Gram-positive bacteria, arbuscular mycorrhizae, and glomalin in small (0.25–1.0 mm) and large (>1 mm) water-stable macro-aggregates were also negatively affected by endophyte infection. Although microbial changes due to endophyte infection were not ubiquitous and overwhelming, they were consistent with previous observations of reduced decomposition of endophyte-infected tall fescue plant litter, which may lead to greater soil C sequestration. Keywords Biolog . Community-level physiological profiles . Microbial diversity . Phospholipid fatty acids . Substrate utilization assay Abbreviations CFU Colony forming unit E− Endophyte free E+ Endophyte infected PLFA Phospholipid fatty acid
K. A. Nichols USDA—Agricultural Research Service, 1701 10th Avenue SW, P.O. Box 459, Mandan, ND 58554, USA
Introduction A. J. Franzluebbers (*) USDA—Agricultural Research Service, 1420 Experiment Station Road, Watkinsville, GA 30677, USA e-mail:
[email protected]
Tall fescue [Lolium arundinaceum (Schreb.) S.J. Darbyshire] is an important grass grown around the world for forage and turf and is considered the most
402
important perennial, cool-season grass for cattle production in the southeastern USA. It is naturally infected with a fungal endophyte, Neotyphodium coenophialum, which resides in the above-ground portions of the plant and produces a variety of alkaloids that have been shown to be toxic when consumed in large quantities by grazing cattle, sheep, and horses (Stuedemann and Hoveland 1988). Endophyte-free (E−) tall fescue pastures can be developed by planting seed that has had the fungus killed during prolonged storage. However, endophytefree pastures have not been widely developed for long-term grazing systems, because stands are not persistent due to reduced grazing tolerance and lower disease and pest resistance (Hoveland 1993). The fungal endophyte symbiosis with tall fescue (E+), therefore, is considered an important component in the agroecological fitness of tall fescue (Clay 1997). One of the positive ecological consequences of endophyte infection can be found in soil. Total soil organic C and N contents were greater under high- than under low-endophyte-infected pastures (Franzluebbers et al. 1999). One reason for the difference in soil organic C and N could be related to greater fitness of E+ tall fescue, in which seedling dry matter production can be greater than under E− tall fescue (Cheplick et al. 1989; Clay 1993). In a 60-week growth study, E+ tall fescue produced greater plant biomass than E− tall fescue, although results varied by soil type (Franzluebbers 2006). Greater plant production could lead to greater C input to soil resulting in greater soil organic C, and subsequent changes in soil microbial community composition. Another reason for the difference in soil organic C and N contents could be due to altered soil microbial dynamics. When 8 to 15 year-old tall fescue pastures had high endophyte infection levels, potential C mineralization per unit of soil organic C was lower than when endophyte infection levels were low (Franzluebbers et al. 1999). At the end of 4 years of seed-harvest management of tall fescue in Kentucky, potential C mineralization per unit of microbial biomass C was lower under E+ than under E− tall fescue (Handayani et al. 2010). Leaf litter from E+ tall fescue decomposed slower than from E− tall fescue throughout a 24-week litter-bag study (Siegrist et al. 2010). Further, E+ ryegrass (Lolium multiflorum Lam.) leaf litter decomposed slower than E− leaf litter in a 12-week outdoor microcosm (Omacini et al. 2004).
Plant Soil (2011) 339:401–412
During a 32-d decomposition experiment, C mineralization and microbial biomass C were slightly inhibited by E+ tall fescue leaves compared with E− leaves, but net N mineralization and microbial biomass N were enhanced by endophyte infection (Franzluebbers and Hill 2005). During a 60-week growth study, Archaea and high G+C (guanine and cytosine) Gram-positive bacteria were suppressed by E+ compared with E− in a clay loam soil, but not in a loamy sand soil (Jenkins et al. 2006). Both greater C input and altered microbial processing of leaf litter and soil organic matter appear to be possible mechanisms for greater soil organic C and N accumulation in long-term E+ than E− tall fescue pastures. Further research is needed to understand the potential changes in soil microbial community structure and function in response to endophyte exposure, both in the short- and long-term. Soil microbial community structure and function can be evaluated with a combination of phospholid fatty acid (PLFA) profiles and substrate utilization (Biolog) (Buyer et al. 1999, 2002; Petersen et al. 2002). Biolog has been effective at characterizing the functional capability of soil organisms to utilize specific C substrates (Garland and Mills 1991; Garland 1996; Buyer et al. 2001). The method is relatively simple, rapid, and easy to use, but has been criticized, because it implicitly assumes that bacteria present are culturable, which may not always be the case (Konopka et al. 1998). Soil PLFA profiles can help isolate specific biomarker groups of organisms without the need to culture organisms (Cavigelli et al. 1995; McMahon et al. 2005). However, PLFA profiles are limited with regards to assessing soil function, since they combine organisms into large groups with overlapping functions. How microbial communities change in soil could alter specific soil processes, such as production of glomalin (an aggregating glue-like glycoprotein) by arbuscular mycorrhizae, development of soil structure, and their resulting impacts on water infiltration, nutrient cycling, and soil organic C sequestration (Treseder and Turner 2007). We hypothesized that Biolog and PLFA profiles could discriminate potential microbial community changes in soil in response to endophyte infection of tall fescue. We wanted to associate potential microbial community changes with potential differences in glomalin (as measured by Bradford-reactive soil protein), since differences in water-stable aggregation and C and N contents of aggregate fractions were
Plant Soil (2011) 339:401–412
previously observed in response to endophyte infection (Franzluebbers 2006) and grassland sampling showed strong relationships among glomalin, soil organic C, total soil N, particulate organic C and N, potentially mineralizable C, and soil microbial biomass C (Franzluebbers et al. 2000). Previously, responses of soil prokaryotic communities to endophyte infection of tall fescue in bulk and rhizosphere soil were determined with fluorescent in situ hybridization (Jenkins et al. 2006). A mesocosm experiment was conducted to determine soil microbial community structure and function in response to short-term exposure of soil to E+ and E− tall fescue plants. Two soil types were evaluated to discern the influence of particle size (and presumed difference in soil porosity) and inherent fertility on changes in soil microbial community structure with time. A semicontrolled experiment [i.e. uniform soil and plant populations with naturally variable environmental conditions (although with irrigation and winter protection)] with multiple evaluations of soil was employed to overcome large inherent spatial variability in soil properties in typical pastures and to help discern temporal changes. The working hypothesis was that E + tall fescue would inhibit a portion of the soil microbial community during the experiment, and eventually lead to enhanced organic C and N storage compared with E− tall fescue, as observed in long-term field studies.
Materials and methods Experimental setup An outdoor mesocosm study consisting of 48 experimental units was conducted from March 2002 until April 2003 near Watkinsville GA (33° 52′ N, 83° 25′ W). Climatic conditions during the experiment were reported in Franzluebbers (2006); average monthly temperature ranged from a low of 5°C at 46 weeks to a high of 26°C at 20 weeks. The experimental design consisted of three randomized replications of a factorial arrangement of soil type (clay loam and loamy sand) and endophyte infection [E− and E+ (common toxic endophyte infection)] placed in four blocks, in which 12 containers (2 soil type × 2 endophyte infection × 3 replications) were sequentially harvested at 8, 20, 36, and 60 weeks of growth. The clay loam contained an average of 33% clay and
403
46% sand and the loamy sand contained an average of 24% clay and 64% sand. Soil was not exposed to tall fescue-endophyte association previously and was of Cecil-Pacolet-Appling series (clayey, kaolinitic, thermic Typic Kanhapludults) collected from a depth of ∼1-m (clay loam) and at the surface of an alluvial wash (loamy sand). Details of the container study can be found in Franzluebbers (2006) with a summary described in the following. Containers were 15 cm in diameter and height and filled with 2.5 kg of the clay loam and 2.7 kg of the loamy sand. Tillers from 2- to 3-year-old ‘Jesup’ tall fescue pastures (E− and E+) were excavated, washed, and five tillers placed in each container. At 8, 20, and 36 weeks of growth, those experimental units not removed for plant and soil analyses had the forage clipped ∼3 cm above the soil and placed on the soil surface to decompose. This protocol allowed some return of above-ground plant material to the surface, but did not closely mimic natural pasture dynamics, in which grazing and dung deposition are the major pathways of above-ground plant input and ungrazed plant material senesces without clipping prematurely. Soil sampling and analyses At each of the four sampling dates, 12 experimental units were removed for analysis. Roots were separated from soil by hand-working contents over a screen with 8-mm openings. Subsamples were kept field-moist at 4°C for Biolog and plate counting analyses and at −20°C for PLFA analysis. The remaining portion of soil was dried at 55°C for 3 days and passed through a screen with 4.75-mm openings and then split into two subsamples—one for water-stable aggregation and glomalin analyses and one for biochemical and physical soil analyses reported in Franzluebbers (2006). Plate counts Soil samples (10 g) were diluted to 10−5 in sterile deionized water. Soil suspension (1 mL) was spread onto 0.1-strength tryptic soy agar plates and incubated at 30°C for 72 h. Plate counts were conducted in triplicate and mean values associated with each experimental unit were subjected to further statistical analysis. Colony forming units were recorded and data were transformed to log 10 values for statistical analyses.
404
Substrate utilization assay (Biolog) Soil samples (10 g) were diluted to 10−3 in sterile deionized water. Soil suspension (150 μL, 10−3 dilution) was added to each well of Biolog GN plates. Plates were incubated for 48 h at 30°C and scanned on a LabSystems Multiskan MS at 590 nm. Average well-color development was recorded as absorbance values ranging from 0 to 2. Substrates were categorized into seven groups for statistical analysis: polymers, carbohydrates, carboxylic acids, amines and amides, amino acids, miscellaneous, and the sum of all other groups (Zak et al. 1994). Phospholipid fatty acid (PLFA) analysis The method was previously described (Buyer et al. 2010). Briefly, lipids were extracted from 5 g of lyophilized soil using a modified Bligh-Dyer extraction. Lipids were fractionated by silica gel chromatography and the phospholipid fraction then trans-esterified in mild alkaline methanol. Resulting fatty acid methyl esters were purified by solid phase extraction chromatography using a 100 mg NH2 column (Phenomenex). Each sample was loaded and eluted in chloroform. After evaporation of chloroform, samples were dissolved in 200 μL of 1:1 hexane:methyl tert-butyl ether and analyzed with an Agilent 6890 gas chromatograph. Fatty acid methyl esters were identified using MIDI software (MIDI Inc., Newark DE), with the eukaryotic method modified to use a split ratio of 1:50. Hexadecanoic acid methyl ester was used as an external standard for quantification. Identification of fatty acids was confirmed for randomly chosen samples using a gas chromatograph-mass spectrometer (Hewlett-Packard 5890 GC, 5970 MSD). Fatty acids were categorized into seven groups for statistical analysis: Gram-positive bacteria (iso and anteiso branched), Gram-negative bacteria (monounsaturated), actinomycetes (10-methyl fatty acids), fungi (18:2 ω6 cis), protozoa (20:3 ω6 cis and 20:4 ω6 cis), arbuscular mycorrhizae (16:1 ω5 cis), and total PLFAs. Glomalin Soil from each experimental unit was separated into the following components for glomalin analysis: water-stable micro-aggregates (0.053–0.25 mm), small water-stable macro-aggregates (0.25–1.0 mm),
Plant Soil (2011) 339:401–412
large water-stable macro-aggregates (>1 mm), particulate organic matter fraction (>0.053 mm), and whole soil. Details of fractionation procedures were described in Franzluebbers (2006). Glomalin was extracted with 100 mM sodium pyrophosphate, pH 9.0 (Wright et al. 2006). All extractions were performed at 121°C for an hour. After each 1-h cycle, samples were centrifuged so that the supernatant could be decanted from the pelleted soil. The extraction procedure was repeated until the extract solution became straw-colored. Decanted solution from each extraction cycle was combined and assayed for glomalin concentration using the Bradfordreactive total protein assay (Rillig 2004; Wright et al. 1996). In some cases, the amount of glomalin in the extract was too dilute and the sample volume was concentrated by evaporation at 70°C. Glomalin data were calculated on a g−1 soil basis, accounting for the fraction of whole soil in the aggregate or particulate fraction. Statistical analyses Analysis of variance was conducted using the general linear model in SAS (Cary NC) following the experimental setup consisting of a factorial arrangement of soil type (clay loam and loamy sand) and endophyte infection (E− and E+) in a completely randomized design with three replications and repeated measurement at four time periods (8, 20, 36, and 60 weeks). Endophyte-infection treatment effects were analyzed as a main (1° of freedom, df) and interactive effect with soil type (1 df) using the replication × soil type × endophyte infection variation as an error term (8 df). Interaction of endophyte infection with sampling period (3 df) was also tested using the residual variation (24 df). Effects were considered significant at p≤0.05. Correlation among microbial variables (reported here) and soil biochemical and physical properties [reported in Franzluebbers (2006)] was conducted on mean values for each soil type × endophyte infection × sampling period (n=16), excluding the initial soil analyses. Associations were considered significant at p≤0.01. Redundancy analysis was conducted using Canoco 4.5 to analyze relationships among PLFAs, treatments (endophyte infection and soil type), and soil properties. Decomposition of variance was conducted as described by ter Braak and Smilauer (2002).
Plant Soil (2011) 339:401–412
405
Microbial community function from the sum of the six substrate groups (i.e. polymers, carbohydrates, carboxylic acids, amines and amides, amino acids, and miscellaneous) varied with soil, endophyte infection, and time (Table 1). Highest summed functional activity was at Week 36 and lowest at Week 60. Across sampling periods and endophyte infection levels, summed functional activity was lower (p
